ChatBot_Intentions_detector.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!conda install -c anaconda nltk --yes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import nltk\n",
    "from tensorflow import keras\n",
    "from nltk.stem.lancaster import LancasterStemmer\n",
    "#stemmer = LancasterStemmer()\n",
    "from nltk.stem import SnowballStemmer\n",
    "stemmer = SnowballStemmer('spanish')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "words=[]\n",
    "documents=[\"crear correo electronico\",\"crear email\",\"nuevo correo electronico\",\"como estas\",\"dime la hora\",\"que hora es\",\"en que hora vivimos\",\"que tiempo hace\",\"prediccion meteorologica\",\"que tal te encuentras\",\"estas bien\",\"que ha dicho el meteorologo\"]\n",
    "classes=[\"crear email\",\"salud\",\"hora\",\"tiempo\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['crear correo electronico',\n",
       " 'crear email',\n",
       " 'nuevo correo electronico',\n",
       " 'como estas',\n",
       " 'dime la hora',\n",
       " 'que hora es',\n",
       " 'en que hora vivimos',\n",
       " 'que tiempo hace',\n",
       " 'prediccion meteorologica',\n",
       " 'que tal te encuentras',\n",
       " 'estas bien',\n",
       " 'que ha dicho el meteorologo']"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "documents"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "document_words=[]\n",
    "for sentence in documents:\n",
    "    w = nltk.word_tokenize(sentence)\n",
    "    # add to our words list\n",
    "    words.extend(w)\n",
    "    document_words.append(w)\n",
    "    # stem and lower each word and remove duplicates\n",
    "words = [stemmer.stem(w.lower()) for w in words]\n",
    "words = sorted(list(set(words)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['bien',\n",
       " 'com',\n",
       " 'corre',\n",
       " 'cre',\n",
       " 'dich',\n",
       " 'dim',\n",
       " 'el',\n",
       " 'electron',\n",
       " 'email',\n",
       " 'en',\n",
       " 'encuentr',\n",
       " 'es',\n",
       " 'estas',\n",
       " 'ha',\n",
       " 'hac',\n",
       " 'hor',\n",
       " 'la',\n",
       " 'meteorolog',\n",
       " 'nuev',\n",
       " 'prediccion',\n",
       " 'que',\n",
       " 'tal',\n",
       " 'te',\n",
       " 'tiemp',\n",
       " 'viv']"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['crear', 'correo', 'electronico'],\n",
       " ['crear', 'email'],\n",
       " ['nuevo', 'correo', 'electronico'],\n",
       " ['como', 'estas'],\n",
       " ['dime', 'la', 'hora'],\n",
       " ['que', 'hora', 'es'],\n",
       " ['en', 'que', 'hora', 'vivimos'],\n",
       " ['que', 'tiempo', 'hace'],\n",
       " ['prediccion', 'meteorologica'],\n",
       " ['que', 'tal', 'te', 'encuentras'],\n",
       " ['estas', 'bien'],\n",
       " ['que', 'ha', 'dicho', 'el', 'meteorologo']]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "document_words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['crear', 'correo', 'electronico']\n",
      "[0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
      "['crear', 'email']\n",
      "[0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
      "['nuevo', 'correo', 'electronico']\n",
      "[0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0]\n",
      "['como', 'estas']\n",
      "[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
      "['dime', 'la', 'hora']\n",
      "[0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]\n",
      "['que', 'hora', 'es']\n",
      "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0]\n",
      "['en', 'que', 'hora', 'vivimos']\n",
      "[0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1]\n",
      "['que', 'tiempo', 'hace']\n",
      "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0]\n",
      "['prediccion', 'meteorologica']\n",
      "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0]\n",
      "['que', 'tal', 'te', 'encuentras']\n",
      "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0]\n",
      "['estas', 'bien']\n",
      "[1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
      "['que', 'ha', 'dicho', 'el', 'meteorologo']\n",
      "[0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0]\n"
     ]
    }
   ],
   "source": [
    "document_bags=[]\n",
    "for sentence in document_words:\n",
    "    print(sentence)\n",
    "    bag=[0] * len(words)\n",
    "    for wds in sentence:\n",
    "        w=stemmer.stem(wds.lower())\n",
    "        for i,wd in enumerate(words):\n",
    "            if wd==w:\n",
    "                bag[i]=1\n",
    "    print(bag)\n",
    "    document_bags.append(bag)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "output=[[1,0,0,0],[1,0,0,0],[1,0,0,0],[0,1,0,0],[0,0,1,0],[0,0,1,0],[0,0,1,0],[0,0,0,1],[0,0,0,1],[0,1,0,0],[0,1,0,0],[0,0,0,1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_x=np.array(document_bags)\n",
    "train_y=np.array(output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 25)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 4)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_3 (Dense)              (None, 10)                260       \n",
      "_________________________________________________________________\n",
      "activation_3 (Activation)    (None, 10)                0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 10)                0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 10)                110       \n",
      "_________________________________________________________________\n",
      "activation_4 (Activation)    (None, 10)                0         \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 10)                0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 4)                 44        \n",
      "_________________________________________________________________\n",
      "activation_5 (Activation)    (None, 4)                 0         \n",
      "=================================================================\n",
      "Total params: 414\n",
      "Trainable params: 414\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Activation, Dropout\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Dense(10, input_dim=train_x.shape[1]))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(10))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(4))\n",
    "model.add(Activation('softmax'))\n",
    "model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# For a multi-class classification problem\n",
    "model.compile(optimizer='rmsprop',\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2.2.4-tf'"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "keras.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:The `nb_epoch` argument in `fit` has been renamed `epochs`.\n",
      "Train on 12 samples\n",
      "Epoch 1/1000\n",
      "12/12 [==============================] - 0s 28ms/sample - loss: 1.6540 - accuracy: 0.2500\n",
      "Epoch 2/1000\n",
      "12/12 [==============================] - 0s 762us/sample - loss: 1.5044 - accuracy: 0.2500\n",
      "Epoch 3/1000\n",
      "12/12 [==============================] - 0s 913us/sample - loss: 1.4494 - accuracy: 0.3333\n",
      "Epoch 4/1000\n",
      "12/12 [==============================] - 0s 783us/sample - loss: 1.1797 - accuracy: 0.5000\n",
      "Epoch 5/1000\n",
      "12/12 [==============================] - 0s 790us/sample - loss: 1.4982 - accuracy: 0.2500\n",
      "Epoch 6/1000\n",
      "12/12 [==============================] - 0s 844us/sample - loss: 1.5638 - accuracy: 0.0833\n",
      "Epoch 7/1000\n",
      "12/12 [==============================] - 0s 899us/sample - loss: 1.3807 - accuracy: 0.2500\n",
      "Epoch 8/1000\n",
      "12/12 [==============================] - 0s 786us/sample - loss: 1.4439 - accuracy: 0.1667\n",
      "Epoch 9/1000\n",
      "12/12 [==============================] - 0s 948us/sample - loss: 1.4164 - accuracy: 0.2500\n",
      "Epoch 10/1000\n",
      "12/12 [==============================] - 0s 857us/sample - loss: 1.5079 - accuracy: 0.1667\n",
      "Epoch 11/1000\n",
      "12/12 [==============================] - 0s 838us/sample - loss: 1.3465 - accuracy: 0.3333\n",
      "Epoch 12/1000\n",
      "12/12 [==============================] - 0s 885us/sample - loss: 1.5221 - accuracy: 0.1667\n",
      "Epoch 13/1000\n",
      "12/12 [==============================] - 0s 913us/sample - loss: 1.6064 - accuracy: 0.0833\n",
      "Epoch 14/1000\n",
      "12/12 [==============================] - 0s 798us/sample - loss: 1.4130 - accuracy: 0.3333\n",
      "Epoch 15/1000\n",
      "12/12 [==============================] - 0s 944us/sample - loss: 1.2924 - accuracy: 0.3333\n",
      "Epoch 16/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.5808 - accuracy: 0.1667\n",
      "Epoch 17/1000\n",
      "12/12 [==============================] - 0s 980us/sample - loss: 1.7561 - accuracy: 0.0833\n",
      "Epoch 18/1000\n",
      "12/12 [==============================] - 0s 996us/sample - loss: 1.3883 - accuracy: 0.2500\n",
      "Epoch 19/1000\n",
      "12/12 [==============================] - 0s 962us/sample - loss: 1.4271 - accuracy: 0.3333\n",
      "Epoch 20/1000\n",
      "12/12 [==============================] - 0s 832us/sample - loss: 1.5439 - accuracy: 0.0000e+00\n",
      "Epoch 21/1000\n",
      "12/12 [==============================] - 0s 852us/sample - loss: 1.3696 - accuracy: 0.3333\n",
      "Epoch 22/1000\n",
      "12/12 [==============================] - 0s 892us/sample - loss: 1.4930 - accuracy: 0.1667\n",
      "Epoch 23/1000\n",
      "12/12 [==============================] - 0s 913us/sample - loss: 1.3724 - accuracy: 0.1667\n",
      "Epoch 24/1000\n",
      "12/12 [==============================] - 0s 788us/sample - loss: 1.6246 - accuracy: 0.0833\n",
      "Epoch 25/1000\n",
      "12/12 [==============================] - 0s 903us/sample - loss: 1.3607 - accuracy: 0.2500\n",
      "Epoch 26/1000\n",
      "12/12 [==============================] - 0s 855us/sample - loss: 1.4422 - accuracy: 0.3333\n",
      "Epoch 27/1000\n",
      "12/12 [==============================] - 0s 868us/sample - loss: 1.4142 - accuracy: 0.4167\n",
      "Epoch 28/1000\n",
      "12/12 [==============================] - 0s 1000us/sample - loss: 1.4137 - accuracy: 0.1667\n",
      "Epoch 29/1000\n",
      "12/12 [==============================] - 0s 953us/sample - loss: 1.4241 - accuracy: 0.3333\n",
      "Epoch 30/1000\n",
      "12/12 [==============================] - 0s 856us/sample - loss: 1.2506 - accuracy: 0.2500\n",
      "Epoch 31/1000\n",
      "12/12 [==============================] - 0s 849us/sample - loss: 1.4978 - accuracy: 0.1667\n",
      "Epoch 32/1000\n",
      "12/12 [==============================] - 0s 926us/sample - loss: 1.4540 - accuracy: 0.2500\n",
      "Epoch 33/1000\n",
      "12/12 [==============================] - 0s 889us/sample - loss: 1.4100 - accuracy: 0.2500\n",
      "Epoch 34/1000\n",
      "12/12 [==============================] - 0s 788us/sample - loss: 1.5742 - accuracy: 0.0000e+00\n",
      "Epoch 35/1000\n",
      "12/12 [==============================] - 0s 815us/sample - loss: 1.2322 - accuracy: 0.6667\n",
      "Epoch 36/1000\n",
      "12/12 [==============================] - 0s 804us/sample - loss: 1.3367 - accuracy: 0.1667\n",
      "Epoch 37/1000\n",
      "12/12 [==============================] - 0s 941us/sample - loss: 1.3422 - accuracy: 0.3333\n",
      "Epoch 38/1000\n",
      "12/12 [==============================] - 0s 843us/sample - loss: 1.3962 - accuracy: 0.2500\n",
      "Epoch 39/1000\n",
      "12/12 [==============================] - 0s 887us/sample - loss: 1.4964 - accuracy: 0.0833\n",
      "Epoch 40/1000\n",
      "12/12 [==============================] - 0s 848us/sample - loss: 1.3018 - accuracy: 0.2500\n",
      "Epoch 41/1000\n",
      "12/12 [==============================] - 0s 860us/sample - loss: 1.1830 - accuracy: 0.4167\n",
      "Epoch 42/1000\n",
      "12/12 [==============================] - 0s 979us/sample - loss: 1.3445 - accuracy: 0.2500\n",
      "Epoch 43/1000\n",
      "12/12 [==============================] - 0s 877us/sample - loss: 1.3674 - accuracy: 0.3333\n",
      "Epoch 44/1000\n",
      "12/12 [==============================] - 0s 824us/sample - loss: 1.3467 - accuracy: 0.2500\n",
      "Epoch 45/1000\n",
      "12/12 [==============================] - 0s 806us/sample - loss: 1.4038 - accuracy: 0.1667\n",
      "Epoch 46/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2295 - accuracy: 0.4167\n",
      "Epoch 47/1000\n",
      "12/12 [==============================] - 0s 852us/sample - loss: 1.4572 - accuracy: 0.3333\n",
      "Epoch 48/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.6015 - accuracy: 0.0833\n",
      "Epoch 49/1000\n",
      "12/12 [==============================] - 0s 903us/sample - loss: 1.3834 - accuracy: 0.5833\n",
      "Epoch 50/1000\n",
      "12/12 [==============================] - 0s 890us/sample - loss: 1.2366 - accuracy: 0.5000\n",
      "Epoch 51/1000\n",
      "12/12 [==============================] - 0s 791us/sample - loss: 1.4297 - accuracy: 0.3333\n",
      "Epoch 52/1000\n",
      "12/12 [==============================] - 0s 911us/sample - loss: 1.2366 - accuracy: 0.4167\n",
      "Epoch 53/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2466 - accuracy: 0.5000\n",
      "Epoch 54/1000\n",
      "12/12 [==============================] - 0s 892us/sample - loss: 1.2916 - accuracy: 0.5000\n",
      "Epoch 55/1000\n",
      "12/12 [==============================] - 0s 871us/sample - loss: 1.2828 - accuracy: 0.2500\n",
      "Epoch 56/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2199 - accuracy: 0.4167\n",
      "Epoch 57/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.3207 - accuracy: 0.6667\n",
      "Epoch 58/1000\n",
      "12/12 [==============================] - 0s 910us/sample - loss: 1.2242 - accuracy: 0.4167\n",
      "Epoch 59/1000\n",
      "12/12 [==============================] - 0s 859us/sample - loss: 1.3847 - accuracy: 0.0833\n",
      "Epoch 60/1000\n",
      "12/12 [==============================] - 0s 936us/sample - loss: 1.2254 - accuracy: 0.3333\n",
      "Epoch 61/1000\n",
      "12/12 [==============================] - 0s 842us/sample - loss: 1.3571 - accuracy: 0.2500\n",
      "Epoch 62/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2865 - accuracy: 0.4167\n",
      "Epoch 63/1000\n",
      "12/12 [==============================] - 0s 814us/sample - loss: 1.2311 - accuracy: 0.3333\n",
      "Epoch 64/1000\n",
      "12/12 [==============================] - 0s 856us/sample - loss: 1.1578 - accuracy: 0.5000\n",
      "Epoch 65/1000\n",
      "12/12 [==============================] - 0s 813us/sample - loss: 1.3604 - accuracy: 0.5000\n",
      "Epoch 66/1000\n",
      "12/12 [==============================] - 0s 920us/sample - loss: 1.1494 - accuracy: 0.5833\n",
      "Epoch 67/1000\n",
      "12/12 [==============================] - 0s 816us/sample - loss: 1.3778 - accuracy: 0.2500\n",
      "Epoch 68/1000\n",
      "12/12 [==============================] - 0s 866us/sample - loss: 1.2356 - accuracy: 0.3333\n",
      "Epoch 69/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 1.2998 - accuracy: 0.2500\n",
      "Epoch 70/1000\n",
      "12/12 [==============================] - 0s 887us/sample - loss: 1.3107 - accuracy: 0.3333\n",
      "Epoch 71/1000\n",
      "12/12 [==============================] - 0s 854us/sample - loss: 1.3284 - accuracy: 0.1667\n",
      "Epoch 72/1000\n",
      "12/12 [==============================] - 0s 873us/sample - loss: 1.2638 - accuracy: 0.4167\n",
      "Epoch 73/1000\n",
      "12/12 [==============================] - 0s 841us/sample - loss: 1.3072 - accuracy: 0.3333\n",
      "Epoch 74/1000\n",
      "12/12 [==============================] - 0s 946us/sample - loss: 1.4164 - accuracy: 0.2500\n",
      "Epoch 75/1000\n",
      "12/12 [==============================] - 0s 873us/sample - loss: 1.1635 - accuracy: 0.4167\n",
      "Epoch 76/1000\n",
      "12/12 [==============================] - 0s 890us/sample - loss: 1.2833 - accuracy: 0.1667\n",
      "Epoch 77/1000\n",
      "12/12 [==============================] - 0s 878us/sample - loss: 1.3147 - accuracy: 0.4167\n",
      "Epoch 78/1000\n",
      "12/12 [==============================] - 0s 971us/sample - loss: 1.2758 - accuracy: 0.4167\n",
      "Epoch 79/1000\n",
      "12/12 [==============================] - 0s 898us/sample - loss: 1.0915 - accuracy: 0.5833\n",
      "Epoch 80/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 1.3761 - accuracy: 0.1667\n",
      "Epoch 81/1000\n",
      "12/12 [==============================] - 0s 988us/sample - loss: 1.2333 - accuracy: 0.2500\n",
      "Epoch 82/1000\n",
      "12/12 [==============================] - 0s 939us/sample - loss: 1.3961 - accuracy: 0.0833\n",
      "Epoch 83/1000\n",
      "12/12 [==============================] - 0s 811us/sample - loss: 1.4042 - accuracy: 0.2500\n",
      "Epoch 84/1000\n",
      "12/12 [==============================] - 0s 925us/sample - loss: 1.3863 - accuracy: 0.4167\n",
      "Epoch 85/1000\n",
      "12/12 [==============================] - 0s 941us/sample - loss: 1.2488 - accuracy: 0.5000\n",
      "Epoch 86/1000\n",
      "12/12 [==============================] - 0s 829us/sample - loss: 1.3268 - accuracy: 0.3333\n",
      "Epoch 87/1000\n",
      "12/12 [==============================] - 0s 941us/sample - loss: 1.1530 - accuracy: 0.4167\n",
      "Epoch 88/1000\n",
      "12/12 [==============================] - 0s 935us/sample - loss: 1.3279 - accuracy: 0.4167\n",
      "Epoch 89/1000\n",
      "12/12 [==============================] - 0s 833us/sample - loss: 1.2452 - accuracy: 0.2500\n",
      "Epoch 90/1000\n",
      "12/12 [==============================] - 0s 885us/sample - loss: 1.2852 - accuracy: 0.1667\n",
      "Epoch 91/1000\n",
      "12/12 [==============================] - 0s 930us/sample - loss: 1.2250 - accuracy: 0.4167\n",
      "Epoch 92/1000\n",
      "12/12 [==============================] - 0s 831us/sample - loss: 1.4163 - accuracy: 0.0833\n",
      "Epoch 93/1000\n",
      "12/12 [==============================] - 0s 902us/sample - loss: 1.2590 - accuracy: 0.5833\n",
      "Epoch 94/1000\n",
      "12/12 [==============================] - 0s 886us/sample - loss: 1.2221 - accuracy: 0.5833\n",
      "Epoch 95/1000\n",
      "12/12 [==============================] - 0s 976us/sample - loss: 1.2922 - accuracy: 0.3333\n",
      "Epoch 96/1000\n",
      "12/12 [==============================] - 0s 855us/sample - loss: 1.2863 - accuracy: 0.3333\n",
      "Epoch 97/1000\n",
      "12/12 [==============================] - 0s 869us/sample - loss: 1.3664 - accuracy: 0.1667\n",
      "Epoch 98/1000\n",
      "12/12 [==============================] - 0s 890us/sample - loss: 1.2115 - accuracy: 0.4167\n",
      "Epoch 99/1000\n",
      "12/12 [==============================] - 0s 828us/sample - loss: 1.3321 - accuracy: 0.1667\n",
      "Epoch 100/1000\n",
      "12/12 [==============================] - 0s 958us/sample - loss: 1.4297 - accuracy: 0.2500\n",
      "Epoch 101/1000\n",
      "12/12 [==============================] - 0s 906us/sample - loss: 1.3215 - accuracy: 0.3333\n",
      "Epoch 102/1000\n",
      "12/12 [==============================] - 0s 768us/sample - loss: 1.2328 - accuracy: 0.4167\n",
      "Epoch 103/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1880 - accuracy: 0.4167\n",
      "Epoch 104/1000\n",
      "12/12 [==============================] - 0s 983us/sample - loss: 1.3055 - accuracy: 0.5000\n",
      "Epoch 105/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.3630 - accuracy: 0.2500\n",
      "Epoch 106/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2771 - accuracy: 0.3333\n",
      "Epoch 107/1000\n",
      "12/12 [==============================] - 0s 860us/sample - loss: 1.1316 - accuracy: 0.5000\n",
      "Epoch 108/1000\n",
      "12/12 [==============================] - 0s 921us/sample - loss: 1.2392 - accuracy: 0.3333\n",
      "Epoch 109/1000\n",
      "12/12 [==============================] - 0s 871us/sample - loss: 1.0668 - accuracy: 0.6667\n",
      "Epoch 110/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2111 - accuracy: 0.4167\n",
      "Epoch 111/1000\n",
      "12/12 [==============================] - 0s 880us/sample - loss: 1.1840 - accuracy: 0.5000\n",
      "Epoch 112/1000\n",
      "12/12 [==============================] - 0s 944us/sample - loss: 1.3307 - accuracy: 0.1667\n",
      "Epoch 113/1000\n",
      "12/12 [==============================] - 0s 983us/sample - loss: 1.1517 - accuracy: 0.5000\n",
      "Epoch 114/1000\n",
      "12/12 [==============================] - 0s 782us/sample - loss: 1.1356 - accuracy: 0.5000\n",
      "Epoch 115/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.4004 - accuracy: 0.3333\n",
      "Epoch 116/1000\n",
      "12/12 [==============================] - 0s 842us/sample - loss: 1.2516 - accuracy: 0.4167\n",
      "Epoch 117/1000\n",
      "12/12 [==============================] - 0s 848us/sample - loss: 1.2248 - accuracy: 0.3333\n",
      "Epoch 118/1000\n",
      "12/12 [==============================] - 0s 854us/sample - loss: 1.4196 - accuracy: 0.2500\n",
      "Epoch 119/1000\n",
      "12/12 [==============================] - 0s 800us/sample - loss: 1.2540 - accuracy: 0.2500\n",
      "Epoch 120/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.3584 - accuracy: 0.2500\n",
      "Epoch 121/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2556 - accuracy: 0.3333\n",
      "Epoch 122/1000\n",
      "12/12 [==============================] - 0s 906us/sample - loss: 1.1156 - accuracy: 0.5000\n",
      "Epoch 123/1000\n",
      "12/12 [==============================] - 0s 918us/sample - loss: 1.1849 - accuracy: 0.4167\n",
      "Epoch 124/1000\n",
      "12/12 [==============================] - 0s 911us/sample - loss: 1.2303 - accuracy: 0.4167\n",
      "Epoch 125/1000\n",
      "12/12 [==============================] - 0s 975us/sample - loss: 1.1981 - accuracy: 0.5000\n",
      "Epoch 126/1000\n",
      "12/12 [==============================] - 0s 869us/sample - loss: 1.1841 - accuracy: 0.5000\n",
      "Epoch 127/1000\n",
      "12/12 [==============================] - 0s 865us/sample - loss: 1.1785 - accuracy: 0.5000\n",
      "Epoch 128/1000\n",
      "12/12 [==============================] - 0s 902us/sample - loss: 1.2351 - accuracy: 0.3333\n",
      "Epoch 129/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1701 - accuracy: 0.5000\n",
      "Epoch 130/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2141 - accuracy: 0.2500\n",
      "Epoch 131/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.3221 - accuracy: 0.5000\n",
      "Epoch 132/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2341 - accuracy: 0.3333\n",
      "Epoch 133/1000\n",
      "12/12 [==============================] - 0s 962us/sample - loss: 1.2350 - accuracy: 0.4167\n",
      "Epoch 134/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1876 - accuracy: 0.4167\n",
      "Epoch 135/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1399 - accuracy: 0.5000\n",
      "Epoch 136/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.3695 - accuracy: 0.3333\n",
      "Epoch 137/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2803 - accuracy: 0.3333\n",
      "Epoch 138/1000\n",
      "12/12 [==============================] - 0s 840us/sample - loss: 1.1065 - accuracy: 0.5000\n",
      "Epoch 139/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 1.0703 - accuracy: 0.6667\n",
      "Epoch 140/1000\n",
      "12/12 [==============================] - 0s 816us/sample - loss: 1.2902 - accuracy: 0.3333\n",
      "Epoch 141/1000\n",
      "12/12 [==============================] - 0s 801us/sample - loss: 1.3811 - accuracy: 0.4167\n",
      "Epoch 142/1000\n",
      "12/12 [==============================] - 0s 864us/sample - loss: 1.1641 - accuracy: 0.4167\n",
      "Epoch 143/1000\n",
      "12/12 [==============================] - 0s 795us/sample - loss: 1.0520 - accuracy: 0.6667\n",
      "Epoch 144/1000\n",
      "12/12 [==============================] - 0s 810us/sample - loss: 1.0789 - accuracy: 0.7500\n",
      "Epoch 145/1000\n",
      "12/12 [==============================] - 0s 900us/sample - loss: 1.0748 - accuracy: 0.6667\n",
      "Epoch 146/1000\n",
      "12/12 [==============================] - 0s 794us/sample - loss: 1.0723 - accuracy: 0.5833\n",
      "Epoch 147/1000\n",
      "12/12 [==============================] - 0s 839us/sample - loss: 1.0738 - accuracy: 0.4167\n",
      "Epoch 148/1000\n",
      "12/12 [==============================] - 0s 910us/sample - loss: 1.2344 - accuracy: 0.3333\n",
      "Epoch 149/1000\n",
      "12/12 [==============================] - 0s 802us/sample - loss: 0.9820 - accuracy: 0.5833\n",
      "Epoch 150/1000\n",
      "12/12 [==============================] - 0s 813us/sample - loss: 1.1258 - accuracy: 0.5000\n",
      "Epoch 151/1000\n",
      "12/12 [==============================] - 0s 853us/sample - loss: 1.1947 - accuracy: 0.4167\n",
      "Epoch 152/1000\n",
      "12/12 [==============================] - 0s 743us/sample - loss: 1.3544 - accuracy: 0.4167\n",
      "Epoch 153/1000\n",
      "12/12 [==============================] - 0s 882us/sample - loss: 1.1506 - accuracy: 0.5000\n",
      "Epoch 154/1000\n",
      "12/12 [==============================] - 0s 809us/sample - loss: 1.1574 - accuracy: 0.4167\n",
      "Epoch 155/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 1.2536 - accuracy: 0.5000\n",
      "Epoch 156/1000\n",
      "12/12 [==============================] - 0s 870us/sample - loss: 1.1743 - accuracy: 0.4167\n",
      "Epoch 157/1000\n",
      "12/12 [==============================] - 0s 823us/sample - loss: 1.2432 - accuracy: 0.4167\n",
      "Epoch 158/1000\n",
      "12/12 [==============================] - 0s 849us/sample - loss: 1.2749 - accuracy: 0.3333\n",
      "Epoch 159/1000\n",
      "12/12 [==============================] - 0s 855us/sample - loss: 1.2140 - accuracy: 0.4167\n",
      "Epoch 160/1000\n",
      "12/12 [==============================] - 0s 793us/sample - loss: 1.1348 - accuracy: 0.7500\n",
      "Epoch 161/1000\n",
      "12/12 [==============================] - 0s 910us/sample - loss: 1.1652 - accuracy: 0.5833\n",
      "Epoch 162/1000\n",
      "12/12 [==============================] - 0s 780us/sample - loss: 1.0382 - accuracy: 0.5000\n",
      "Epoch 163/1000\n",
      "12/12 [==============================] - 0s 794us/sample - loss: 1.1637 - accuracy: 0.5000\n",
      "Epoch 164/1000\n",
      "12/12 [==============================] - 0s 917us/sample - loss: 1.0375 - accuracy: 0.5833\n",
      "Epoch 165/1000\n",
      "12/12 [==============================] - 0s 921us/sample - loss: 0.9029 - accuracy: 0.7500\n",
      "Epoch 166/1000\n",
      "12/12 [==============================] - 0s 793us/sample - loss: 1.3142 - accuracy: 0.3333\n",
      "Epoch 167/1000\n",
      "12/12 [==============================] - 0s 824us/sample - loss: 1.2937 - accuracy: 0.2500\n",
      "Epoch 168/1000\n",
      "12/12 [==============================] - 0s 776us/sample - loss: 1.0104 - accuracy: 0.6667\n",
      "Epoch 169/1000\n",
      "12/12 [==============================] - 0s 854us/sample - loss: 1.0511 - accuracy: 0.5833\n",
      "Epoch 170/1000\n",
      "12/12 [==============================] - 0s 827us/sample - loss: 1.1291 - accuracy: 0.5000\n",
      "Epoch 171/1000\n",
      "12/12 [==============================] - 0s 775us/sample - loss: 1.0756 - accuracy: 0.5000\n",
      "Epoch 172/1000\n",
      "12/12 [==============================] - 0s 857us/sample - loss: 1.2676 - accuracy: 0.4167\n",
      "Epoch 173/1000\n",
      "12/12 [==============================] - 0s 844us/sample - loss: 1.0653 - accuracy: 0.5000\n",
      "Epoch 174/1000\n",
      "12/12 [==============================] - 0s 831us/sample - loss: 1.3096 - accuracy: 0.5000\n",
      "Epoch 175/1000\n",
      "12/12 [==============================] - 0s 972us/sample - loss: 0.9182 - accuracy: 0.6667\n",
      "Epoch 176/1000\n",
      "12/12 [==============================] - 0s 823us/sample - loss: 1.0356 - accuracy: 0.5833\n",
      "Epoch 177/1000\n",
      "12/12 [==============================] - 0s 761us/sample - loss: 1.3273 - accuracy: 0.3333\n",
      "Epoch 178/1000\n",
      "12/12 [==============================] - 0s 940us/sample - loss: 1.1303 - accuracy: 0.5000\n",
      "Epoch 179/1000\n",
      "12/12 [==============================] - 0s 868us/sample - loss: 1.2465 - accuracy: 0.3333\n",
      "Epoch 180/1000\n",
      "12/12 [==============================] - 0s 806us/sample - loss: 0.9869 - accuracy: 0.6667\n",
      "Epoch 181/1000\n",
      "12/12 [==============================] - 0s 882us/sample - loss: 0.9438 - accuracy: 0.7500\n",
      "Epoch 182/1000\n",
      "12/12 [==============================] - 0s 925us/sample - loss: 1.3629 - accuracy: 0.2500\n",
      "Epoch 183/1000\n",
      "12/12 [==============================] - 0s 884us/sample - loss: 1.3535 - accuracy: 0.3333\n",
      "Epoch 184/1000\n",
      "12/12 [==============================] - 0s 883us/sample - loss: 1.0546 - accuracy: 0.5833\n",
      "Epoch 185/1000\n",
      "12/12 [==============================] - 0s 937us/sample - loss: 1.0777 - accuracy: 0.5833\n",
      "Epoch 186/1000\n",
      "12/12 [==============================] - 0s 951us/sample - loss: 1.1619 - accuracy: 0.5833\n",
      "Epoch 187/1000\n",
      "12/12 [==============================] - 0s 957us/sample - loss: 0.9870 - accuracy: 0.5833\n",
      "Epoch 188/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2135 - accuracy: 0.3333\n",
      "Epoch 189/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1118 - accuracy: 0.4167\n",
      "Epoch 190/1000\n",
      "12/12 [==============================] - 0s 974us/sample - loss: 1.1746 - accuracy: 0.4167\n",
      "Epoch 191/1000\n",
      "12/12 [==============================] - 0s 981us/sample - loss: 1.3167 - accuracy: 0.2500\n",
      "Epoch 192/1000\n",
      "12/12 [==============================] - 0s 929us/sample - loss: 1.3129 - accuracy: 0.3333\n",
      "Epoch 193/1000\n",
      "12/12 [==============================] - 0s 928us/sample - loss: 1.1267 - accuracy: 0.4167\n",
      "Epoch 194/1000\n",
      "12/12 [==============================] - 0s 795us/sample - loss: 1.1357 - accuracy: 0.6667\n",
      "Epoch 195/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.9847 - accuracy: 0.7500\n",
      "Epoch 196/1000\n",
      "12/12 [==============================] - 0s 872us/sample - loss: 1.2722 - accuracy: 0.3333\n",
      "Epoch 197/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 1.3111 - accuracy: 0.5000\n",
      "Epoch 198/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1988 - accuracy: 0.5000\n",
      "Epoch 199/1000\n",
      "12/12 [==============================] - 0s 810us/sample - loss: 1.2054 - accuracy: 0.5000\n",
      "Epoch 200/1000\n",
      "12/12 [==============================] - 0s 889us/sample - loss: 1.2231 - accuracy: 0.4167\n",
      "Epoch 201/1000\n",
      "12/12 [==============================] - 0s 844us/sample - loss: 1.2600 - accuracy: 0.3333\n",
      "Epoch 202/1000\n",
      "12/12 [==============================] - 0s 838us/sample - loss: 1.2975 - accuracy: 0.3333\n",
      "Epoch 203/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.0071 - accuracy: 0.5833\n",
      "Epoch 204/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.8492 - accuracy: 0.6667\n",
      "Epoch 205/1000\n",
      "12/12 [==============================] - 0s 992us/sample - loss: 0.9621 - accuracy: 0.5833\n",
      "Epoch 206/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2435 - accuracy: 0.3333\n",
      "Epoch 207/1000\n",
      "12/12 [==============================] - 0s 892us/sample - loss: 1.0875 - accuracy: 0.5833\n",
      "Epoch 208/1000\n",
      "12/12 [==============================] - 0s 842us/sample - loss: 0.9834 - accuracy: 0.8333\n",
      "Epoch 209/1000\n",
      "12/12 [==============================] - 0s 851us/sample - loss: 1.1806 - accuracy: 0.5000\n",
      "Epoch 210/1000\n",
      "12/12 [==============================] - 0s 822us/sample - loss: 1.0319 - accuracy: 0.5833\n",
      "Epoch 211/1000\n",
      "12/12 [==============================] - 0s 934us/sample - loss: 1.2158 - accuracy: 0.5000\n",
      "Epoch 212/1000\n",
      "12/12 [==============================] - 0s 936us/sample - loss: 1.3413 - accuracy: 0.3333\n",
      "Epoch 213/1000\n",
      "12/12 [==============================] - 0s 821us/sample - loss: 1.0556 - accuracy: 0.5000\n",
      "Epoch 214/1000\n",
      "12/12 [==============================] - 0s 827us/sample - loss: 0.8635 - accuracy: 0.7500\n",
      "Epoch 215/1000\n",
      "12/12 [==============================] - 0s 949us/sample - loss: 1.1272 - accuracy: 0.4167\n",
      "Epoch 216/1000\n",
      "12/12 [==============================] - 0s 807us/sample - loss: 1.2331 - accuracy: 0.3333\n",
      "Epoch 217/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 1.0962 - accuracy: 0.5833\n",
      "Epoch 218/1000\n",
      "12/12 [==============================] - 0s 828us/sample - loss: 1.1687 - accuracy: 0.5000\n",
      "Epoch 219/1000\n",
      "12/12 [==============================] - 0s 748us/sample - loss: 1.1326 - accuracy: 0.5000\n",
      "Epoch 220/1000\n",
      "12/12 [==============================] - 0s 945us/sample - loss: 0.9691 - accuracy: 0.6667\n",
      "Epoch 221/1000\n",
      "12/12 [==============================] - 0s 810us/sample - loss: 1.0995 - accuracy: 0.5000\n",
      "Epoch 222/1000\n",
      "12/12 [==============================] - 0s 762us/sample - loss: 1.1123 - accuracy: 0.7500\n",
      "Epoch 223/1000\n",
      "12/12 [==============================] - 0s 875us/sample - loss: 1.2015 - accuracy: 0.4167\n",
      "Epoch 224/1000\n",
      "12/12 [==============================] - 0s 937us/sample - loss: 1.0635 - accuracy: 0.5000\n",
      "Epoch 225/1000\n",
      "12/12 [==============================] - 0s 802us/sample - loss: 1.0409 - accuracy: 0.5000\n",
      "Epoch 226/1000\n",
      "12/12 [==============================] - 0s 832us/sample - loss: 1.0738 - accuracy: 0.5000\n",
      "Epoch 227/1000\n",
      "12/12 [==============================] - 0s 811us/sample - loss: 0.9918 - accuracy: 0.6667\n",
      "Epoch 228/1000\n",
      "12/12 [==============================] - 0s 762us/sample - loss: 1.1347 - accuracy: 0.6667\n",
      "Epoch 229/1000\n",
      "12/12 [==============================] - 0s 848us/sample - loss: 1.1467 - accuracy: 0.3333\n",
      "Epoch 230/1000\n",
      "12/12 [==============================] - 0s 835us/sample - loss: 1.1516 - accuracy: 0.3333\n",
      "Epoch 231/1000\n",
      "12/12 [==============================] - 0s 919us/sample - loss: 1.0845 - accuracy: 0.5000\n",
      "Epoch 232/1000\n",
      "12/12 [==============================] - 0s 887us/sample - loss: 1.0062 - accuracy: 0.6667\n",
      "Epoch 233/1000\n",
      "12/12 [==============================] - 0s 882us/sample - loss: 0.8556 - accuracy: 0.8333\n",
      "Epoch 234/1000\n",
      "12/12 [==============================] - 0s 819us/sample - loss: 0.8776 - accuracy: 0.6667\n",
      "Epoch 235/1000\n",
      "12/12 [==============================] - 0s 800us/sample - loss: 1.1537 - accuracy: 0.5833\n",
      "Epoch 236/1000\n",
      "12/12 [==============================] - 0s 801us/sample - loss: 1.1364 - accuracy: 0.5000\n",
      "Epoch 237/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1494 - accuracy: 0.5833\n",
      "Epoch 238/1000\n",
      "12/12 [==============================] - 0s 794us/sample - loss: 0.7815 - accuracy: 0.9167\n",
      "Epoch 239/1000\n",
      "12/12 [==============================] - 0s 808us/sample - loss: 1.2013 - accuracy: 0.5000\n",
      "Epoch 240/1000\n",
      "12/12 [==============================] - 0s 967us/sample - loss: 1.0559 - accuracy: 0.7500\n",
      "Epoch 241/1000\n",
      "12/12 [==============================] - 0s 823us/sample - loss: 1.2648 - accuracy: 0.3333\n",
      "Epoch 242/1000\n",
      "12/12 [==============================] - 0s 822us/sample - loss: 1.1448 - accuracy: 0.6667\n",
      "Epoch 243/1000\n",
      "12/12 [==============================] - 0s 881us/sample - loss: 1.1793 - accuracy: 0.5000\n",
      "Epoch 244/1000\n",
      "12/12 [==============================] - 0s 788us/sample - loss: 1.0197 - accuracy: 0.8333\n",
      "Epoch 245/1000\n",
      "12/12 [==============================] - 0s 803us/sample - loss: 1.1745 - accuracy: 0.4167\n",
      "Epoch 246/1000\n",
      "12/12 [==============================] - 0s 802us/sample - loss: 1.2849 - accuracy: 0.3333\n",
      "Epoch 247/1000\n",
      "12/12 [==============================] - 0s 714us/sample - loss: 1.0715 - accuracy: 0.5833\n",
      "Epoch 248/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 0.7448 - accuracy: 0.6667\n",
      "Epoch 249/1000\n",
      "12/12 [==============================] - 0s 912us/sample - loss: 1.0347 - accuracy: 0.6667\n",
      "Epoch 250/1000\n",
      "12/12 [==============================] - 0s 816us/sample - loss: 1.1229 - accuracy: 0.3333\n",
      "Epoch 251/1000\n",
      "12/12 [==============================] - 0s 789us/sample - loss: 1.3763 - accuracy: 0.3333\n",
      "Epoch 252/1000\n",
      "12/12 [==============================] - 0s 808us/sample - loss: 1.0137 - accuracy: 0.6667\n",
      "Epoch 253/1000\n",
      "12/12 [==============================] - 0s 909us/sample - loss: 0.9842 - accuracy: 0.5833\n",
      "Epoch 254/1000\n",
      "12/12 [==============================] - 0s 794us/sample - loss: 1.0979 - accuracy: 0.5000\n",
      "Epoch 255/1000\n",
      "12/12 [==============================] - 0s 852us/sample - loss: 1.1373 - accuracy: 0.4167\n",
      "Epoch 256/1000\n",
      "12/12 [==============================] - 0s 794us/sample - loss: 1.0048 - accuracy: 0.5833\n",
      "Epoch 257/1000\n",
      "12/12 [==============================] - 0s 948us/sample - loss: 1.0183 - accuracy: 0.5000\n",
      "Epoch 258/1000\n",
      "12/12 [==============================] - 0s 873us/sample - loss: 0.9812 - accuracy: 0.6667\n",
      "Epoch 259/1000\n",
      "12/12 [==============================] - 0s 801us/sample - loss: 1.0546 - accuracy: 0.5000\n",
      "Epoch 260/1000\n",
      "12/12 [==============================] - 0s 814us/sample - loss: 1.1347 - accuracy: 0.5000\n",
      "Epoch 261/1000\n",
      "12/12 [==============================] - 0s 777us/sample - loss: 0.9001 - accuracy: 0.5000\n",
      "Epoch 262/1000\n",
      "12/12 [==============================] - 0s 908us/sample - loss: 0.9428 - accuracy: 0.6667\n",
      "Epoch 263/1000\n",
      "12/12 [==============================] - 0s 744us/sample - loss: 1.1891 - accuracy: 0.4167\n",
      "Epoch 264/1000\n",
      "12/12 [==============================] - 0s 860us/sample - loss: 0.9727 - accuracy: 0.6667\n",
      "Epoch 265/1000\n",
      "12/12 [==============================] - 0s 798us/sample - loss: 1.1837 - accuracy: 0.4167\n",
      "Epoch 266/1000\n",
      "12/12 [==============================] - 0s 829us/sample - loss: 0.9837 - accuracy: 0.5833\n",
      "Epoch 267/1000\n",
      "12/12 [==============================] - 0s 839us/sample - loss: 0.9106 - accuracy: 0.5000\n",
      "Epoch 268/1000\n",
      "12/12 [==============================] - 0s 880us/sample - loss: 0.9695 - accuracy: 0.7500\n",
      "Epoch 269/1000\n",
      "12/12 [==============================] - 0s 736us/sample - loss: 1.1272 - accuracy: 0.4167\n",
      "Epoch 270/1000\n",
      "12/12 [==============================] - 0s 784us/sample - loss: 1.3386 - accuracy: 0.2500\n",
      "Epoch 271/1000\n",
      "12/12 [==============================] - 0s 796us/sample - loss: 1.0036 - accuracy: 0.7500\n",
      "Epoch 272/1000\n",
      "12/12 [==============================] - 0s 926us/sample - loss: 1.0661 - accuracy: 0.5000\n",
      "Epoch 273/1000\n",
      "12/12 [==============================] - 0s 801us/sample - loss: 1.2640 - accuracy: 0.3333\n",
      "Epoch 274/1000\n",
      "12/12 [==============================] - 0s 849us/sample - loss: 1.2283 - accuracy: 0.4167\n",
      "Epoch 275/1000\n",
      "12/12 [==============================] - 0s 825us/sample - loss: 0.6986 - accuracy: 0.8333\n",
      "Epoch 276/1000\n",
      "12/12 [==============================] - 0s 872us/sample - loss: 0.9629 - accuracy: 0.6667\n",
      "Epoch 277/1000\n",
      "12/12 [==============================] - 0s 854us/sample - loss: 0.8604 - accuracy: 0.8333\n",
      "Epoch 278/1000\n",
      "12/12 [==============================] - 0s 790us/sample - loss: 1.1688 - accuracy: 0.4167\n",
      "Epoch 279/1000\n",
      "12/12 [==============================] - 0s 820us/sample - loss: 0.7558 - accuracy: 0.7500\n",
      "Epoch 280/1000\n",
      "12/12 [==============================] - 0s 851us/sample - loss: 1.1087 - accuracy: 0.4167\n",
      "Epoch 281/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1032 - accuracy: 0.5833\n",
      "Epoch 282/1000\n",
      "12/12 [==============================] - 0s 801us/sample - loss: 1.0507 - accuracy: 0.5833\n",
      "Epoch 283/1000\n",
      "12/12 [==============================] - 0s 763us/sample - loss: 0.6483 - accuracy: 0.7500\n",
      "Epoch 284/1000\n",
      "12/12 [==============================] - 0s 772us/sample - loss: 1.0436 - accuracy: 0.5833\n",
      "Epoch 285/1000\n",
      "12/12 [==============================] - 0s 781us/sample - loss: 1.2573 - accuracy: 0.2500\n",
      "Epoch 286/1000\n",
      "12/12 [==============================] - 0s 866us/sample - loss: 0.8216 - accuracy: 0.6667\n",
      "Epoch 287/1000\n",
      "12/12 [==============================] - 0s 932us/sample - loss: 1.1226 - accuracy: 0.5000\n",
      "Epoch 288/1000\n",
      "12/12 [==============================] - 0s 835us/sample - loss: 0.9984 - accuracy: 0.6667\n",
      "Epoch 289/1000\n",
      "12/12 [==============================] - 0s 877us/sample - loss: 0.6718 - accuracy: 0.9167\n",
      "Epoch 290/1000\n",
      "12/12 [==============================] - 0s 839us/sample - loss: 1.0840 - accuracy: 0.3333\n",
      "Epoch 291/1000\n",
      "12/12 [==============================] - 0s 793us/sample - loss: 0.9782 - accuracy: 0.4167\n",
      "Epoch 292/1000\n",
      "12/12 [==============================] - 0s 824us/sample - loss: 1.1283 - accuracy: 0.5833\n",
      "Epoch 293/1000\n",
      "12/12 [==============================] - 0s 735us/sample - loss: 0.7291 - accuracy: 0.7500\n",
      "Epoch 294/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.8429 - accuracy: 0.6667\n",
      "Epoch 295/1000\n",
      "12/12 [==============================] - 0s 849us/sample - loss: 1.1526 - accuracy: 0.5833\n",
      "Epoch 296/1000\n",
      "12/12 [==============================] - 0s 820us/sample - loss: 0.8226 - accuracy: 0.7500\n",
      "Epoch 297/1000\n",
      "12/12 [==============================] - 0s 958us/sample - loss: 0.8266 - accuracy: 0.6667\n",
      "Epoch 298/1000\n",
      "12/12 [==============================] - 0s 893us/sample - loss: 0.7955 - accuracy: 0.6667\n",
      "Epoch 299/1000\n",
      "12/12 [==============================] - 0s 719us/sample - loss: 0.8630 - accuracy: 0.6667\n",
      "Epoch 300/1000\n",
      "12/12 [==============================] - 0s 812us/sample - loss: 0.8205 - accuracy: 0.7500\n",
      "Epoch 301/1000\n",
      "12/12 [==============================] - 0s 853us/sample - loss: 1.0203 - accuracy: 0.5000\n",
      "Epoch 302/1000\n",
      "12/12 [==============================] - 0s 819us/sample - loss: 1.0975 - accuracy: 0.4167\n",
      "Epoch 303/1000\n",
      "12/12 [==============================] - 0s 764us/sample - loss: 0.9163 - accuracy: 0.6667\n",
      "Epoch 304/1000\n",
      "12/12 [==============================] - 0s 800us/sample - loss: 1.0352 - accuracy: 0.5833\n",
      "Epoch 305/1000\n",
      "12/12 [==============================] - 0s 812us/sample - loss: 1.0904 - accuracy: 0.5000\n",
      "Epoch 306/1000\n",
      "12/12 [==============================] - 0s 814us/sample - loss: 0.9794 - accuracy: 0.5833\n",
      "Epoch 307/1000\n",
      "12/12 [==============================] - ETA: 0s - loss: 1.0062 - accuracy: 0.66 - 0s 788us/sample - loss: 0.9566 - accuracy: 0.5833\n",
      "Epoch 308/1000\n",
      "12/12 [==============================] - 0s 811us/sample - loss: 1.1595 - accuracy: 0.5833\n",
      "Epoch 309/1000\n",
      "12/12 [==============================] - 0s 818us/sample - loss: 1.1216 - accuracy: 0.5000\n",
      "Epoch 310/1000\n",
      "12/12 [==============================] - 0s 836us/sample - loss: 1.1174 - accuracy: 0.4167\n",
      "Epoch 311/1000\n",
      "12/12 [==============================] - 0s 843us/sample - loss: 0.9644 - accuracy: 0.5000\n",
      "Epoch 312/1000\n",
      "12/12 [==============================] - 0s 808us/sample - loss: 1.0260 - accuracy: 0.5833\n",
      "Epoch 313/1000\n",
      "12/12 [==============================] - 0s 813us/sample - loss: 0.9452 - accuracy: 0.6667\n",
      "Epoch 314/1000\n",
      "12/12 [==============================] - 0s 849us/sample - loss: 0.6497 - accuracy: 0.8333\n",
      "Epoch 315/1000\n",
      "12/12 [==============================] - 0s 784us/sample - loss: 1.0160 - accuracy: 0.5000\n",
      "Epoch 316/1000\n",
      "12/12 [==============================] - 0s 890us/sample - loss: 1.0525 - accuracy: 0.4167\n",
      "Epoch 317/1000\n",
      "12/12 [==============================] - 0s 852us/sample - loss: 0.8360 - accuracy: 0.6667\n",
      "Epoch 318/1000\n",
      "12/12 [==============================] - 0s 783us/sample - loss: 1.1084 - accuracy: 0.4167\n",
      "Epoch 319/1000\n",
      "12/12 [==============================] - 0s 928us/sample - loss: 1.0886 - accuracy: 0.5000\n",
      "Epoch 320/1000\n",
      "12/12 [==============================] - 0s 934us/sample - loss: 0.8327 - accuracy: 0.5833\n",
      "Epoch 321/1000\n",
      "12/12 [==============================] - 0s 786us/sample - loss: 1.3197 - accuracy: 0.3333\n",
      "Epoch 322/1000\n",
      "12/12 [==============================] - 0s 868us/sample - loss: 0.9784 - accuracy: 0.5000\n",
      "Epoch 323/1000\n",
      "12/12 [==============================] - 0s 858us/sample - loss: 0.7058 - accuracy: 0.8333\n",
      "Epoch 324/1000\n",
      "12/12 [==============================] - 0s 791us/sample - loss: 1.0201 - accuracy: 0.6667\n",
      "Epoch 325/1000\n",
      "12/12 [==============================] - 0s 860us/sample - loss: 0.8917 - accuracy: 0.7500\n",
      "Epoch 326/1000\n",
      "12/12 [==============================] - 0s 768us/sample - loss: 1.0341 - accuracy: 0.5833\n",
      "Epoch 327/1000\n",
      "12/12 [==============================] - 0s 899us/sample - loss: 1.2136 - accuracy: 0.2500\n",
      "Epoch 328/1000\n",
      "12/12 [==============================] - 0s 820us/sample - loss: 1.1165 - accuracy: 0.5833\n",
      "Epoch 329/1000\n",
      "12/12 [==============================] - 0s 815us/sample - loss: 0.9824 - accuracy: 0.5833\n",
      "Epoch 330/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 0.7197 - accuracy: 0.8333\n",
      "Epoch 331/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 1.0408 - accuracy: 0.6667\n",
      "Epoch 332/1000\n",
      "12/12 [==============================] - 0s 786us/sample - loss: 0.8641 - accuracy: 0.5833\n",
      "Epoch 333/1000\n",
      "12/12 [==============================] - 0s 785us/sample - loss: 0.8831 - accuracy: 0.5000\n",
      "Epoch 334/1000\n",
      "12/12 [==============================] - 0s 900us/sample - loss: 0.9758 - accuracy: 0.6667\n",
      "Epoch 335/1000\n",
      "12/12 [==============================] - 0s 870us/sample - loss: 0.7620 - accuracy: 0.8333\n",
      "Epoch 336/1000\n",
      "12/12 [==============================] - 0s 892us/sample - loss: 1.2141 - accuracy: 0.4167\n",
      "Epoch 337/1000\n",
      "12/12 [==============================] - 0s 904us/sample - loss: 1.1138 - accuracy: 0.6667\n",
      "Epoch 338/1000\n",
      "12/12 [==============================] - 0s 855us/sample - loss: 0.7464 - accuracy: 0.6667\n",
      "Epoch 339/1000\n",
      "12/12 [==============================] - 0s 944us/sample - loss: 1.3179 - accuracy: 0.2500\n",
      "Epoch 340/1000\n",
      "12/12 [==============================] - 0s 974us/sample - loss: 0.8776 - accuracy: 0.5000\n",
      "Epoch 341/1000\n",
      "12/12 [==============================] - 0s 989us/sample - loss: 1.0406 - accuracy: 0.5833\n",
      "Epoch 342/1000\n",
      "12/12 [==============================] - 0s 892us/sample - loss: 1.0163 - accuracy: 0.5000\n",
      "Epoch 343/1000\n",
      "12/12 [==============================] - 0s 978us/sample - loss: 1.0435 - accuracy: 0.5000\n",
      "Epoch 344/1000\n",
      "12/12 [==============================] - 0s 909us/sample - loss: 0.9126 - accuracy: 0.6667\n",
      "Epoch 345/1000\n",
      "12/12 [==============================] - 0s 873us/sample - loss: 0.9765 - accuracy: 0.4167\n",
      "Epoch 346/1000\n",
      "12/12 [==============================] - 0s 941us/sample - loss: 1.1123 - accuracy: 0.4167\n",
      "Epoch 347/1000\n",
      "12/12 [==============================] - 0s 913us/sample - loss: 0.9329 - accuracy: 0.5833\n",
      "Epoch 348/1000\n",
      "12/12 [==============================] - 0s 856us/sample - loss: 0.8588 - accuracy: 0.7500\n",
      "Epoch 349/1000\n",
      "12/12 [==============================] - 0s 923us/sample - loss: 1.1646 - accuracy: 0.4167\n",
      "Epoch 350/1000\n",
      "12/12 [==============================] - 0s 934us/sample - loss: 0.7898 - accuracy: 0.7500\n",
      "Epoch 351/1000\n",
      "12/12 [==============================] - 0s 955us/sample - loss: 0.8701 - accuracy: 0.6667\n",
      "Epoch 352/1000\n",
      "12/12 [==============================] - 0s 987us/sample - loss: 0.9855 - accuracy: 0.5833\n",
      "Epoch 353/1000\n",
      "12/12 [==============================] - 0s 865us/sample - loss: 0.9286 - accuracy: 0.7500\n",
      "Epoch 354/1000\n",
      "12/12 [==============================] - 0s 943us/sample - loss: 1.2445 - accuracy: 0.2500\n",
      "Epoch 355/1000\n",
      "12/12 [==============================] - 0s 932us/sample - loss: 0.6937 - accuracy: 0.7500\n",
      "Epoch 356/1000\n",
      "12/12 [==============================] - 0s 795us/sample - loss: 0.8172 - accuracy: 0.6667\n",
      "Epoch 357/1000\n",
      "12/12 [==============================] - 0s 874us/sample - loss: 0.9179 - accuracy: 0.6667\n",
      "Epoch 358/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 1.0106 - accuracy: 0.5833\n",
      "Epoch 359/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 0.7918 - accuracy: 0.6667\n",
      "Epoch 360/1000\n",
      "12/12 [==============================] - 0s 840us/sample - loss: 1.1504 - accuracy: 0.5000\n",
      "Epoch 361/1000\n",
      "12/12 [==============================] - 0s 818us/sample - loss: 0.8770 - accuracy: 0.6667\n",
      "Epoch 362/1000\n",
      "12/12 [==============================] - 0s 851us/sample - loss: 0.9053 - accuracy: 0.6667\n",
      "Epoch 363/1000\n",
      "12/12 [==============================] - 0s 853us/sample - loss: 0.8908 - accuracy: 0.5000\n",
      "Epoch 364/1000\n",
      "12/12 [==============================] - 0s 849us/sample - loss: 0.7496 - accuracy: 0.6667\n",
      "Epoch 365/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.0762 - accuracy: 0.5000\n",
      "Epoch 366/1000\n",
      "12/12 [==============================] - 0s 940us/sample - loss: 1.1184 - accuracy: 0.5000\n",
      "Epoch 367/1000\n",
      "12/12 [==============================] - 0s 847us/sample - loss: 1.0972 - accuracy: 0.4167\n",
      "Epoch 368/1000\n",
      "12/12 [==============================] - 0s 965us/sample - loss: 0.8474 - accuracy: 0.6667\n",
      "Epoch 369/1000\n",
      "12/12 [==============================] - 0s 925us/sample - loss: 0.7522 - accuracy: 0.6667\n",
      "Epoch 370/1000\n",
      "12/12 [==============================] - 0s 947us/sample - loss: 0.8958 - accuracy: 0.5000\n",
      "Epoch 371/1000\n",
      "12/12 [==============================] - 0s 980us/sample - loss: 0.8435 - accuracy: 0.5833\n",
      "Epoch 372/1000\n",
      "12/12 [==============================] - 0s 950us/sample - loss: 0.8889 - accuracy: 0.5833\n",
      "Epoch 373/1000\n",
      "12/12 [==============================] - 0s 944us/sample - loss: 0.8803 - accuracy: 0.5833\n",
      "Epoch 374/1000\n",
      "12/12 [==============================] - 0s 984us/sample - loss: 0.7024 - accuracy: 0.8333\n",
      "Epoch 375/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 0.4624 - accuracy: 0.8333\n",
      "Epoch 376/1000\n",
      "12/12 [==============================] - 0s 952us/sample - loss: 1.0218 - accuracy: 0.4167\n",
      "Epoch 377/1000\n",
      "12/12 [==============================] - 0s 943us/sample - loss: 0.9426 - accuracy: 0.6667\n",
      "Epoch 378/1000\n",
      "12/12 [==============================] - 0s 935us/sample - loss: 0.9987 - accuracy: 0.4167\n",
      "Epoch 379/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.1668 - accuracy: 0.2500\n",
      "Epoch 380/1000\n",
      "12/12 [==============================] - 0s 837us/sample - loss: 1.0372 - accuracy: 0.5000\n",
      "Epoch 381/1000\n",
      "12/12 [==============================] - 0s 753us/sample - loss: 0.8805 - accuracy: 0.5833\n",
      "Epoch 382/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.5592 - accuracy: 0.9167\n",
      "Epoch 383/1000\n",
      "12/12 [==============================] - 0s 838us/sample - loss: 0.5266 - accuracy: 0.8333\n",
      "Epoch 384/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 0.9375 - accuracy: 0.5000\n",
      "Epoch 385/1000\n",
      "12/12 [==============================] - 0s 966us/sample - loss: 0.8796 - accuracy: 0.5833\n",
      "Epoch 386/1000\n",
      "12/12 [==============================] - 0s 829us/sample - loss: 1.1346 - accuracy: 0.5833\n",
      "Epoch 387/1000\n",
      "12/12 [==============================] - 0s 789us/sample - loss: 1.0098 - accuracy: 0.5833\n",
      "Epoch 388/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2590 - accuracy: 0.1667\n",
      "Epoch 389/1000\n",
      "12/12 [==============================] - 0s 996us/sample - loss: 0.8434 - accuracy: 0.8333\n",
      "Epoch 390/1000\n",
      "12/12 [==============================] - 0s 793us/sample - loss: 0.8930 - accuracy: 0.6667\n",
      "Epoch 391/1000\n",
      "12/12 [==============================] - 0s 960us/sample - loss: 0.9227 - accuracy: 0.5000\n",
      "Epoch 392/1000\n",
      "12/12 [==============================] - 0s 846us/sample - loss: 1.1527 - accuracy: 0.2500\n",
      "Epoch 393/1000\n",
      "12/12 [==============================] - 0s 787us/sample - loss: 0.9763 - accuracy: 0.4167\n",
      "Epoch 394/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.7904 - accuracy: 0.7500\n",
      "Epoch 395/1000\n",
      "12/12 [==============================] - 0s 834us/sample - loss: 1.2135 - accuracy: 0.4167\n",
      "Epoch 396/1000\n",
      "12/12 [==============================] - 0s 771us/sample - loss: 1.0541 - accuracy: 0.5833\n",
      "Epoch 397/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.9928 - accuracy: 0.5000\n",
      "Epoch 398/1000\n",
      "12/12 [==============================] - 0s 995us/sample - loss: 0.7552 - accuracy: 0.8333\n",
      "Epoch 399/1000\n",
      "12/12 [==============================] - 0s 956us/sample - loss: 0.6085 - accuracy: 0.7500\n",
      "Epoch 400/1000\n",
      "12/12 [==============================] - 0s 864us/sample - loss: 0.7600 - accuracy: 0.6667\n",
      "Epoch 401/1000\n",
      "12/12 [==============================] - 0s 865us/sample - loss: 0.7592 - accuracy: 0.6667\n",
      "Epoch 402/1000\n",
      "12/12 [==============================] - 0s 828us/sample - loss: 0.8965 - accuracy: 0.6667\n",
      "Epoch 403/1000\n",
      "12/12 [==============================] - 0s 774us/sample - loss: 1.1265 - accuracy: 0.5833\n",
      "Epoch 404/1000\n",
      "12/12 [==============================] - 0s 951us/sample - loss: 1.2179 - accuracy: 0.3333\n",
      "Epoch 405/1000\n",
      "12/12 [==============================] - 0s 836us/sample - loss: 1.3140 - accuracy: 0.4167\n",
      "Epoch 406/1000\n",
      "12/12 [==============================] - 0s 799us/sample - loss: 1.0386 - accuracy: 0.5000\n",
      "Epoch 407/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 1.2316 - accuracy: 0.5000\n",
      "Epoch 408/1000\n",
      "12/12 [==============================] - 0s 845us/sample - loss: 0.9010 - accuracy: 0.5000\n",
      "Epoch 409/1000\n",
      "12/12 [==============================] - 0s 813us/sample - loss: 0.8677 - accuracy: 0.7500\n",
      "Epoch 410/1000\n",
      "12/12 [==============================] - 0s 944us/sample - loss: 0.8936 - accuracy: 0.5833\n",
      "Epoch 411/1000\n",
      "12/12 [==============================] - 0s 916us/sample - loss: 0.7453 - accuracy: 0.7500\n",
      "Epoch 412/1000\n",
      "12/12 [==============================] - 0s 866us/sample - loss: 0.9084 - accuracy: 0.6667\n",
      "Epoch 413/1000\n",
      "12/12 [==============================] - 0s 989us/sample - loss: 0.4840 - accuracy: 0.8333\n",
      "Epoch 414/1000\n",
      "12/12 [==============================] - 0s 989us/sample - loss: 0.6375 - accuracy: 0.7500\n",
      "Epoch 415/1000\n",
      "12/12 [==============================] - 0s 850us/sample - loss: 0.5908 - accuracy: 0.7500\n",
      "Epoch 416/1000\n",
      "12/12 [==============================] - 0s 989us/sample - loss: 0.9958 - accuracy: 0.5833\n",
      "Epoch 417/1000\n",
      "12/12 [==============================] - 0s 850us/sample - loss: 1.0363 - accuracy: 0.5833\n",
      "Epoch 418/1000\n",
      "12/12 [==============================] - 0s 860us/sample - loss: 0.9280 - accuracy: 0.6667\n",
      "Epoch 419/1000\n",
      "12/12 [==============================] - 0s 970us/sample - loss: 0.9170 - accuracy: 0.7500\n",
      "Epoch 420/1000\n",
      "12/12 [==============================] - 0s 921us/sample - loss: 1.1094 - accuracy: 0.4167\n",
      "Epoch 421/1000\n",
      "12/12 [==============================] - 0s 851us/sample - loss: 0.9423 - accuracy: 0.5000\n",
      "Epoch 422/1000\n",
      "12/12 [==============================] - 0s 881us/sample - loss: 1.1517 - accuracy: 0.4167\n",
      "Epoch 423/1000\n",
      "12/12 [==============================] - 0s 848us/sample - loss: 1.0474 - accuracy: 0.5833\n",
      "Epoch 424/1000\n",
      "12/12 [==============================] - 0s 790us/sample - loss: 0.9194 - accuracy: 0.5000\n",
      "Epoch 425/1000\n",
      "12/12 [==============================] - 0s 990us/sample - loss: 0.8801 - accuracy: 0.6667\n",
      "Epoch 426/1000\n",
      "12/12 [==============================] - 0s 813us/sample - loss: 0.5804 - accuracy: 0.9167\n",
      "Epoch 427/1000\n",
      "12/12 [==============================] - 0s 785us/sample - loss: 0.8111 - accuracy: 0.5833\n",
      "Epoch 428/1000\n",
      "12/12 [==============================] - 0s 908us/sample - loss: 0.5557 - accuracy: 0.9167\n",
      "Epoch 429/1000\n",
      "12/12 [==============================] - 0s 835us/sample - loss: 1.2159 - accuracy: 0.4167\n",
      "Epoch 430/1000\n",
      "12/12 [==============================] - 0s 767us/sample - loss: 1.0732 - accuracy: 0.5000\n",
      "Epoch 431/1000\n",
      "12/12 [==============================] - 0s 879us/sample - loss: 0.8205 - accuracy: 0.5000\n",
      "Epoch 432/1000\n",
      "12/12 [==============================] - 0s 896us/sample - loss: 0.8588 - accuracy: 0.8333\n",
      "Epoch 433/1000\n",
      "12/12 [==============================] - 0s 820us/sample - loss: 1.0468 - accuracy: 0.5833\n",
      "Epoch 434/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.4400 - accuracy: 0.8333\n",
      "Epoch 435/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 0.7904 - accuracy: 0.6667\n",
      "Epoch 436/1000\n",
      "12/12 [==============================] - 0s 791us/sample - loss: 1.0448 - accuracy: 0.4167\n",
      "Epoch 437/1000\n",
      "12/12 [==============================] - 0s 935us/sample - loss: 0.7623 - accuracy: 0.6667\n",
      "Epoch 438/1000\n",
      "12/12 [==============================] - 0s 918us/sample - loss: 0.6737 - accuracy: 0.8333\n",
      "Epoch 439/1000\n",
      "12/12 [==============================] - 0s 833us/sample - loss: 0.6461 - accuracy: 0.8333\n",
      "Epoch 440/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.8550 - accuracy: 0.5833\n",
      "Epoch 441/1000\n",
      "12/12 [==============================] - 0s 778us/sample - loss: 0.8742 - accuracy: 0.6667\n",
      "Epoch 442/1000\n",
      "12/12 [==============================] - 0s 881us/sample - loss: 0.8849 - accuracy: 0.5833\n",
      "Epoch 443/1000\n",
      "12/12 [==============================] - 0s 940us/sample - loss: 1.0094 - accuracy: 0.4167\n",
      "Epoch 444/1000\n",
      "12/12 [==============================] - 0s 828us/sample - loss: 0.9187 - accuracy: 0.5000\n",
      "Epoch 445/1000\n",
      "12/12 [==============================] - 0s 881us/sample - loss: 0.5331 - accuracy: 0.8333\n",
      "Epoch 446/1000\n",
      "12/12 [==============================] - 0s 889us/sample - loss: 0.9818 - accuracy: 0.6667\n",
      "Epoch 447/1000\n",
      "12/12 [==============================] - 0s 819us/sample - loss: 0.8576 - accuracy: 0.5833\n",
      "Epoch 448/1000\n",
      "12/12 [==============================] - 0s 842us/sample - loss: 0.9108 - accuracy: 0.6667\n",
      "Epoch 449/1000\n",
      "12/12 [==============================] - 0s 846us/sample - loss: 0.5790 - accuracy: 0.7500\n",
      "Epoch 450/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 0.7151 - accuracy: 0.7500\n",
      "Epoch 451/1000\n",
      "12/12 [==============================] - 0s 842us/sample - loss: 0.5078 - accuracy: 0.9167\n",
      "Epoch 452/1000\n",
      "12/12 [==============================] - 0s 912us/sample - loss: 0.7227 - accuracy: 0.6667\n",
      "Epoch 453/1000\n",
      "12/12 [==============================] - 0s 817us/sample - loss: 1.1325 - accuracy: 0.5833\n",
      "Epoch 454/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 0.8699 - accuracy: 0.5833\n",
      "Epoch 455/1000\n",
      "12/12 [==============================] - 0s 762us/sample - loss: 0.6915 - accuracy: 0.7500\n",
      "Epoch 456/1000\n",
      "12/12 [==============================] - 0s 826us/sample - loss: 0.7908 - accuracy: 0.5833\n",
      "Epoch 457/1000\n",
      "12/12 [==============================] - 0s 795us/sample - loss: 0.6940 - accuracy: 0.7500\n",
      "Epoch 458/1000\n",
      "12/12 [==============================] - 0s 777us/sample - loss: 0.6992 - accuracy: 0.6667\n",
      "Epoch 459/1000\n",
      "12/12 [==============================] - 0s 885us/sample - loss: 0.7746 - accuracy: 0.6667\n",
      "Epoch 460/1000\n",
      "12/12 [==============================] - 0s 791us/sample - loss: 0.6763 - accuracy: 0.7500\n",
      "Epoch 461/1000\n",
      "12/12 [==============================] - 0s 860us/sample - loss: 0.9419 - accuracy: 0.5833\n",
      "Epoch 462/1000\n",
      "12/12 [==============================] - 0s 888us/sample - loss: 0.9020 - accuracy: 0.5833\n",
      "Epoch 463/1000\n",
      "12/12 [==============================] - 0s 773us/sample - loss: 0.6595 - accuracy: 0.7500\n",
      "Epoch 464/1000\n",
      "12/12 [==============================] - 0s 863us/sample - loss: 0.8402 - accuracy: 0.5833\n",
      "Epoch 465/1000\n",
      "12/12 [==============================] - 0s 757us/sample - loss: 0.6680 - accuracy: 0.7500\n",
      "Epoch 466/1000\n",
      "12/12 [==============================] - 0s 904us/sample - loss: 0.8296 - accuracy: 0.8333\n",
      "Epoch 467/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 0.7514 - accuracy: 0.7500\n",
      "Epoch 468/1000\n",
      "12/12 [==============================] - 0s 749us/sample - loss: 1.0087 - accuracy: 0.6667\n",
      "Epoch 469/1000\n",
      "12/12 [==============================] - 0s 828us/sample - loss: 0.6545 - accuracy: 0.6667\n",
      "Epoch 470/1000\n",
      "12/12 [==============================] - 0s 964us/sample - loss: 0.9057 - accuracy: 0.6667\n",
      "Epoch 471/1000\n",
      "12/12 [==============================] - 0s 725us/sample - loss: 0.8448 - accuracy: 0.7500\n",
      "Epoch 472/1000\n",
      "12/12 [==============================] - 0s 867us/sample - loss: 0.7895 - accuracy: 0.5833\n",
      "Epoch 473/1000\n",
      "12/12 [==============================] - 0s 833us/sample - loss: 0.7520 - accuracy: 0.7500\n",
      "Epoch 474/1000\n",
      "12/12 [==============================] - 0s 773us/sample - loss: 0.8724 - accuracy: 0.6667\n",
      "Epoch 475/1000\n",
      "12/12 [==============================] - 0s 808us/sample - loss: 0.7269 - accuracy: 0.7500\n",
      "Epoch 476/1000\n",
      "12/12 [==============================] - 0s 848us/sample - loss: 0.6793 - accuracy: 0.7500\n",
      "Epoch 477/1000\n",
      "12/12 [==============================] - 0s 787us/sample - loss: 1.0375 - accuracy: 0.3333\n",
      "Epoch 478/1000\n",
      "12/12 [==============================] - 0s 865us/sample - loss: 0.8869 - accuracy: 0.6667\n",
      "Epoch 479/1000\n",
      "12/12 [==============================] - 0s 797us/sample - loss: 0.8313 - accuracy: 0.5833\n",
      "Epoch 480/1000\n",
      "12/12 [==============================] - 0s 865us/sample - loss: 0.9045 - accuracy: 0.6667\n",
      "Epoch 481/1000\n",
      "12/12 [==============================] - 0s 969us/sample - loss: 0.7358 - accuracy: 0.6667\n",
      "Epoch 482/1000\n",
      "12/12 [==============================] - 0s 759us/sample - loss: 0.8213 - accuracy: 0.7500\n",
      "Epoch 483/1000\n",
      "12/12 [==============================] - 0s 824us/sample - loss: 1.0795 - accuracy: 0.5000\n",
      "Epoch 484/1000\n",
      "12/12 [==============================] - 0s 780us/sample - loss: 1.2380 - accuracy: 0.4167\n",
      "Epoch 485/1000\n",
      "12/12 [==============================] - 0s 770us/sample - loss: 0.8873 - accuracy: 0.6667\n",
      "Epoch 486/1000\n",
      "12/12 [==============================] - 0s 886us/sample - loss: 1.0373 - accuracy: 0.5833\n",
      "Epoch 487/1000\n",
      "12/12 [==============================] - 0s 798us/sample - loss: 0.5667 - accuracy: 0.7500\n",
      "Epoch 488/1000\n",
      "12/12 [==============================] - 0s 794us/sample - loss: 0.6418 - accuracy: 0.8333\n",
      "Epoch 489/1000\n",
      "12/12 [==============================] - 0s 795us/sample - loss: 0.5608 - accuracy: 0.8333\n",
      "Epoch 490/1000\n",
      "12/12 [==============================] - 0s 736us/sample - loss: 1.0349 - accuracy: 0.4167\n",
      "Epoch 491/1000\n",
      "12/12 [==============================] - 0s 772us/sample - loss: 0.7373 - accuracy: 0.6667\n",
      "Epoch 492/1000\n",
      "12/12 [==============================] - 0s 713us/sample - loss: 0.7274 - accuracy: 0.6667\n",
      "Epoch 493/1000\n",
      "12/12 [==============================] - 0s 779us/sample - loss: 0.4871 - accuracy: 0.8333\n",
      "Epoch 494/1000\n",
      "12/12 [==============================] - 0s 989us/sample - loss: 0.6403 - accuracy: 0.7500\n",
      "Epoch 495/1000\n",
      "12/12 [==============================] - 0s 944us/sample - loss: 0.8317 - accuracy: 0.7500\n",
      "Epoch 496/1000\n",
      "12/12 [==============================] - 0s 874us/sample - loss: 0.5354 - accuracy: 0.7500\n",
      "Epoch 497/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 0.4375 - accuracy: 0.8333\n",
      "Epoch 498/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.9119 - accuracy: 0.5833\n",
      "Epoch 499/1000\n",
      "12/12 [==============================] - 0s 789us/sample - loss: 0.6479 - accuracy: 0.6667\n",
      "Epoch 500/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 0.6966 - accuracy: 0.7500\n",
      "Epoch 501/1000\n",
      "12/12 [==============================] - 0s 780us/sample - loss: 0.5726 - accuracy: 0.7500\n",
      "Epoch 502/1000\n",
      "12/12 [==============================] - 0s 799us/sample - loss: 1.1703 - accuracy: 0.2500\n",
      "Epoch 503/1000\n",
      "12/12 [==============================] - 0s 691us/sample - loss: 0.9520 - accuracy: 0.5833\n",
      "Epoch 504/1000\n",
      "12/12 [==============================] - 0s 787us/sample - loss: 0.4582 - accuracy: 0.9167\n",
      "Epoch 505/1000\n",
      "12/12 [==============================] - 0s 747us/sample - loss: 0.8672 - accuracy: 0.5000\n",
      "Epoch 506/1000\n",
      "12/12 [==============================] - 0s 945us/sample - loss: 0.8340 - accuracy: 0.6667\n",
      "Epoch 507/1000\n",
      "12/12 [==============================] - 0s 919us/sample - loss: 0.9232 - accuracy: 0.5833\n",
      "Epoch 508/1000\n",
      "12/12 [==============================] - 0s 948us/sample - loss: 0.8516 - accuracy: 0.5833\n",
      "Epoch 509/1000\n",
      "12/12 [==============================] - 0s 937us/sample - loss: 0.4404 - accuracy: 1.0000\n",
      "Epoch 510/1000\n",
      "12/12 [==============================] - 0s 976us/sample - loss: 0.4340 - accuracy: 1.0000\n",
      "Epoch 511/1000\n",
      "12/12 [==============================] - 0s 912us/sample - loss: 0.5356 - accuracy: 0.9167\n",
      "Epoch 512/1000\n",
      "12/12 [==============================] - 0s 855us/sample - loss: 1.0856 - accuracy: 0.5000\n",
      "Epoch 513/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 0.9492 - accuracy: 0.5833\n",
      "Epoch 514/1000\n",
      "12/12 [==============================] - 0s 851us/sample - loss: 0.9260 - accuracy: 0.6667\n",
      "Epoch 515/1000\n",
      "12/12 [==============================] - 0s 783us/sample - loss: 0.7032 - accuracy: 0.6667\n",
      "Epoch 516/1000\n",
      "12/12 [==============================] - 0s 764us/sample - loss: 0.9483 - accuracy: 0.5833\n",
      "Epoch 517/1000\n",
      "12/12 [==============================] - 0s 702us/sample - loss: 0.8991 - accuracy: 0.6667\n",
      "Epoch 518/1000\n",
      "12/12 [==============================] - 0s 880us/sample - loss: 0.9303 - accuracy: 0.5000\n",
      "Epoch 519/1000\n",
      "12/12 [==============================] - 0s 834us/sample - loss: 0.6523 - accuracy: 0.6667\n",
      "Epoch 520/1000\n",
      "12/12 [==============================] - 0s 740us/sample - loss: 0.7078 - accuracy: 0.8333\n",
      "Epoch 521/1000\n",
      "12/12 [==============================] - 0s 784us/sample - loss: 0.7069 - accuracy: 0.6667\n",
      "Epoch 522/1000\n",
      "12/12 [==============================] - 0s 770us/sample - loss: 0.5102 - accuracy: 0.8333\n",
      "Epoch 523/1000\n",
      "12/12 [==============================] - 0s 728us/sample - loss: 0.5554 - accuracy: 0.7500\n",
      "Epoch 524/1000\n",
      "12/12 [==============================] - 0s 805us/sample - loss: 0.8807 - accuracy: 0.6667\n",
      "Epoch 525/1000\n",
      "12/12 [==============================] - 0s 773us/sample - loss: 1.2188 - accuracy: 0.2500\n",
      "Epoch 526/1000\n",
      "12/12 [==============================] - 0s 827us/sample - loss: 0.5751 - accuracy: 0.8333\n",
      "Epoch 527/1000\n",
      "12/12 [==============================] - 0s 805us/sample - loss: 0.9456 - accuracy: 0.5833\n",
      "Epoch 528/1000\n",
      "12/12 [==============================] - 0s 782us/sample - loss: 0.8287 - accuracy: 0.5000\n",
      "Epoch 529/1000\n",
      "12/12 [==============================] - 0s 804us/sample - loss: 0.7748 - accuracy: 0.6667\n",
      "Epoch 530/1000\n",
      "12/12 [==============================] - 0s 710us/sample - loss: 0.6556 - accuracy: 0.8333\n",
      "Epoch 531/1000\n",
      "12/12 [==============================] - 0s 719us/sample - loss: 0.6300 - accuracy: 0.8333\n",
      "Epoch 532/1000\n",
      "12/12 [==============================] - 0s 770us/sample - loss: 0.7727 - accuracy: 0.5833\n",
      "Epoch 533/1000\n",
      "12/12 [==============================] - 0s 764us/sample - loss: 0.7233 - accuracy: 0.6667\n",
      "Epoch 534/1000\n",
      "12/12 [==============================] - 0s 791us/sample - loss: 1.0195 - accuracy: 0.5000\n",
      "Epoch 535/1000\n",
      "12/12 [==============================] - 0s 804us/sample - loss: 0.8451 - accuracy: 0.5833\n",
      "Epoch 536/1000\n",
      "12/12 [==============================] - 0s 807us/sample - loss: 0.6951 - accuracy: 0.8333\n",
      "Epoch 537/1000\n",
      "12/12 [==============================] - 0s 863us/sample - loss: 0.6781 - accuracy: 0.6667\n",
      "Epoch 538/1000\n",
      "12/12 [==============================] - 0s 915us/sample - loss: 0.6780 - accuracy: 0.8333\n",
      "Epoch 539/1000\n",
      "12/12 [==============================] - 0s 815us/sample - loss: 0.9849 - accuracy: 0.6667\n",
      "Epoch 540/1000\n",
      "12/12 [==============================] - 0s 671us/sample - loss: 0.6353 - accuracy: 0.8333\n",
      "Epoch 541/1000\n",
      "12/12 [==============================] - 0s 804us/sample - loss: 0.6458 - accuracy: 0.6667\n",
      "Epoch 542/1000\n",
      "12/12 [==============================] - 0s 809us/sample - loss: 1.0272 - accuracy: 0.5000\n",
      "Epoch 543/1000\n",
      "12/12 [==============================] - 0s 736us/sample - loss: 0.7318 - accuracy: 0.6667\n",
      "Epoch 544/1000\n",
      "12/12 [==============================] - 0s 701us/sample - loss: 0.6176 - accuracy: 0.7500\n",
      "Epoch 545/1000\n",
      "12/12 [==============================] - 0s 810us/sample - loss: 1.1197 - accuracy: 0.5833\n",
      "Epoch 546/1000\n",
      "12/12 [==============================] - 0s 773us/sample - loss: 0.8001 - accuracy: 0.5833\n",
      "Epoch 547/1000\n",
      "12/12 [==============================] - 0s 766us/sample - loss: 0.5502 - accuracy: 0.9167\n",
      "Epoch 548/1000\n",
      "12/12 [==============================] - 0s 818us/sample - loss: 0.5461 - accuracy: 0.8333\n",
      "Epoch 549/1000\n",
      "12/12 [==============================] - 0s 753us/sample - loss: 0.5811 - accuracy: 0.8333\n",
      "Epoch 550/1000\n",
      "12/12 [==============================] - 0s 721us/sample - loss: 0.6707 - accuracy: 0.7500\n",
      "Epoch 551/1000\n",
      "12/12 [==============================] - 0s 789us/sample - loss: 0.8409 - accuracy: 0.5833\n",
      "Epoch 552/1000\n",
      "12/12 [==============================] - 0s 980us/sample - loss: 0.7890 - accuracy: 0.7500\n",
      "Epoch 553/1000\n",
      "12/12 [==============================] - 0s 828us/sample - loss: 0.8645 - accuracy: 0.6667\n",
      "Epoch 554/1000\n",
      "12/12 [==============================] - 0s 833us/sample - loss: 0.8575 - accuracy: 0.5833\n",
      "Epoch 555/1000\n",
      "12/12 [==============================] - 0s 970us/sample - loss: 0.5986 - accuracy: 0.5833\n",
      "Epoch 556/1000\n",
      "12/12 [==============================] - 0s 885us/sample - loss: 0.6518 - accuracy: 0.6667\n",
      "Epoch 557/1000\n",
      "12/12 [==============================] - 0s 970us/sample - loss: 0.6340 - accuracy: 0.7500\n",
      "Epoch 558/1000\n",
      "12/12 [==============================] - 0s 914us/sample - loss: 0.7575 - accuracy: 0.6667\n",
      "Epoch 559/1000\n",
      "12/12 [==============================] - 0s 802us/sample - loss: 0.8710 - accuracy: 0.5833\n",
      "Epoch 560/1000\n",
      "12/12 [==============================] - 0s 973us/sample - loss: 0.8823 - accuracy: 0.5833\n",
      "Epoch 561/1000\n",
      "12/12 [==============================] - 0s 855us/sample - loss: 0.6473 - accuracy: 0.7500\n",
      "Epoch 562/1000\n",
      "12/12 [==============================] - 0s 903us/sample - loss: 0.8373 - accuracy: 0.6667\n",
      "Epoch 563/1000\n",
      "12/12 [==============================] - 0s 919us/sample - loss: 0.6705 - accuracy: 0.7500\n",
      "Epoch 564/1000\n",
      "12/12 [==============================] - 0s 885us/sample - loss: 0.4778 - accuracy: 0.7500\n",
      "Epoch 565/1000\n",
      "12/12 [==============================] - 0s 862us/sample - loss: 0.8946 - accuracy: 0.6667\n",
      "Epoch 566/1000\n",
      "12/12 [==============================] - 0s 961us/sample - loss: 0.7281 - accuracy: 0.5833\n",
      "Epoch 567/1000\n",
      "12/12 [==============================] - 0s 899us/sample - loss: 0.4250 - accuracy: 0.9167\n",
      "Epoch 568/1000\n",
      "12/12 [==============================] - 0s 919us/sample - loss: 0.7948 - accuracy: 0.5833\n",
      "Epoch 569/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 0.5440 - accuracy: 0.9167\n",
      "Epoch 570/1000\n",
      "12/12 [==============================] - 0s 901us/sample - loss: 0.5517 - accuracy: 0.9167\n",
      "Epoch 571/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 0.4677 - accuracy: 0.9167\n",
      "Epoch 572/1000\n",
      "12/12 [==============================] - 0s 921us/sample - loss: 0.7903 - accuracy: 0.6667\n",
      "Epoch 573/1000\n",
      "12/12 [==============================] - 0s 887us/sample - loss: 0.8442 - accuracy: 0.5000\n",
      "Epoch 574/1000\n",
      "12/12 [==============================] - 0s 854us/sample - loss: 0.5344 - accuracy: 0.7500\n",
      "Epoch 575/1000\n",
      "12/12 [==============================] - 0s 934us/sample - loss: 0.8205 - accuracy: 0.7500\n",
      "Epoch 576/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 0.5675 - accuracy: 0.7500\n",
      "Epoch 577/1000\n",
      "12/12 [==============================] - 0s 868us/sample - loss: 0.4544 - accuracy: 0.9167\n",
      "Epoch 578/1000\n",
      "12/12 [==============================] - 0s 890us/sample - loss: 0.7794 - accuracy: 0.7500\n",
      "Epoch 579/1000\n",
      "12/12 [==============================] - 0s 850us/sample - loss: 0.9523 - accuracy: 0.5000\n",
      "Epoch 580/1000\n",
      "12/12 [==============================] - 0s 912us/sample - loss: 0.9127 - accuracy: 0.5833\n",
      "Epoch 581/1000\n",
      "12/12 [==============================] - 0s 938us/sample - loss: 0.8940 - accuracy: 0.6667\n",
      "Epoch 582/1000\n",
      "12/12 [==============================] - 0s 920us/sample - loss: 0.4992 - accuracy: 0.8333\n",
      "Epoch 583/1000\n",
      "12/12 [==============================] - 0s 920us/sample - loss: 0.8501 - accuracy: 0.5833\n",
      "Epoch 584/1000\n",
      "12/12 [==============================] - 0s 915us/sample - loss: 0.9450 - accuracy: 0.5833\n",
      "Epoch 585/1000\n",
      "12/12 [==============================] - 0s 905us/sample - loss: 0.7673 - accuracy: 0.8333\n",
      "Epoch 586/1000\n",
      "12/12 [==============================] - 0s 965us/sample - loss: 0.6532 - accuracy: 0.6667\n",
      "Epoch 587/1000\n",
      "12/12 [==============================] - 0s 941us/sample - loss: 0.7996 - accuracy: 0.6667\n",
      "Epoch 588/1000\n",
      "12/12 [==============================] - 0s 853us/sample - loss: 0.7132 - accuracy: 0.5833\n",
      "Epoch 589/1000\n",
      "12/12 [==============================] - 0s 885us/sample - loss: 0.7045 - accuracy: 0.7500\n",
      "Epoch 590/1000\n",
      "12/12 [==============================] - 0s 899us/sample - loss: 1.0277 - accuracy: 0.5833\n",
      "Epoch 591/1000\n",
      "12/12 [==============================] - 0s 819us/sample - loss: 0.5642 - accuracy: 0.8333\n",
      "Epoch 592/1000\n",
      "12/12 [==============================] - 0s 905us/sample - loss: 0.4345 - accuracy: 0.9167\n",
      "Epoch 593/1000\n",
      "12/12 [==============================] - 0s 924us/sample - loss: 1.1970 - accuracy: 0.5000\n",
      "Epoch 594/1000\n",
      "12/12 [==============================] - 0s 838us/sample - loss: 0.6882 - accuracy: 0.6667\n",
      "Epoch 595/1000\n",
      "12/12 [==============================] - 0s 896us/sample - loss: 0.8204 - accuracy: 0.7500\n",
      "Epoch 596/1000\n",
      "12/12 [==============================] - 0s 890us/sample - loss: 0.9592 - accuracy: 0.5833\n",
      "Epoch 597/1000\n",
      "12/12 [==============================] - 0s 894us/sample - loss: 1.0526 - accuracy: 0.4167\n",
      "Epoch 598/1000\n",
      "12/12 [==============================] - 0s 905us/sample - loss: 1.0275 - accuracy: 0.5000\n",
      "Epoch 599/1000\n",
      "12/12 [==============================] - 0s 953us/sample - loss: 0.6362 - accuracy: 0.9167\n",
      "Epoch 600/1000\n",
      "12/12 [==============================] - 0s 818us/sample - loss: 0.7068 - accuracy: 0.6667\n",
      "Epoch 601/1000\n",
      "12/12 [==============================] - 0s 966us/sample - loss: 1.0259 - accuracy: 0.4167\n",
      "Epoch 602/1000\n",
      "12/12 [==============================] - 0s 917us/sample - loss: 0.6924 - accuracy: 0.7500\n",
      "Epoch 603/1000\n",
      "12/12 [==============================] - 0s 823us/sample - loss: 0.4267 - accuracy: 0.9167\n",
      "Epoch 604/1000\n",
      "12/12 [==============================] - 0s 860us/sample - loss: 0.8228 - accuracy: 0.6667\n",
      "Epoch 605/1000\n",
      "12/12 [==============================] - 0s 924us/sample - loss: 0.8432 - accuracy: 0.5833\n",
      "Epoch 606/1000\n",
      "12/12 [==============================] - 0s 832us/sample - loss: 0.6788 - accuracy: 0.6667\n",
      "Epoch 607/1000\n",
      "12/12 [==============================] - 0s 852us/sample - loss: 0.9957 - accuracy: 0.6667\n",
      "Epoch 608/1000\n",
      "12/12 [==============================] - 0s 698us/sample - loss: 0.9428 - accuracy: 0.5000\n",
      "Epoch 609/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 0.6982 - accuracy: 0.6667\n",
      "Epoch 610/1000\n",
      "12/12 [==============================] - 0s 772us/sample - loss: 0.9426 - accuracy: 0.5833\n",
      "Epoch 611/1000\n",
      "12/12 [==============================] - 0s 771us/sample - loss: 1.1554 - accuracy: 0.3333\n",
      "Epoch 612/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 0.6973 - accuracy: 0.6667\n",
      "Epoch 613/1000\n",
      "12/12 [==============================] - 0s 716us/sample - loss: 1.2879 - accuracy: 0.2500\n",
      "Epoch 614/1000\n",
      "12/12 [==============================] - 0s 791us/sample - loss: 0.7526 - accuracy: 0.6667\n",
      "Epoch 615/1000\n",
      "12/12 [==============================] - 0s 800us/sample - loss: 0.6842 - accuracy: 0.7500\n",
      "Epoch 616/1000\n",
      "12/12 [==============================] - 0s 705us/sample - loss: 1.0444 - accuracy: 0.5833\n",
      "Epoch 617/1000\n",
      "12/12 [==============================] - 0s 783us/sample - loss: 0.5396 - accuracy: 0.8333\n",
      "Epoch 618/1000\n",
      "12/12 [==============================] - 0s 768us/sample - loss: 0.9642 - accuracy: 0.4167\n",
      "Epoch 619/1000\n",
      "12/12 [==============================] - 0s 704us/sample - loss: 0.4507 - accuracy: 0.8333\n",
      "Epoch 620/1000\n",
      "12/12 [==============================] - 0s 858us/sample - loss: 0.5284 - accuracy: 0.8333\n",
      "Epoch 621/1000\n",
      "12/12 [==============================] - 0s 798us/sample - loss: 0.9742 - accuracy: 0.5833\n",
      "Epoch 622/1000\n",
      "12/12 [==============================] - 0s 775us/sample - loss: 1.0360 - accuracy: 0.4167\n",
      "Epoch 623/1000\n",
      "12/12 [==============================] - 0s 744us/sample - loss: 0.6236 - accuracy: 0.6667\n",
      "Epoch 624/1000\n",
      "12/12 [==============================] - 0s 793us/sample - loss: 1.0728 - accuracy: 0.5833\n",
      "Epoch 625/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 1.0678 - accuracy: 0.5000\n",
      "Epoch 626/1000\n",
      "12/12 [==============================] - 0s 706us/sample - loss: 0.6699 - accuracy: 0.7500\n",
      "Epoch 627/1000\n",
      "12/12 [==============================] - 0s 786us/sample - loss: 0.5904 - accuracy: 0.5833\n",
      "Epoch 628/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 0.5456 - accuracy: 0.8333\n",
      "Epoch 629/1000\n",
      "12/12 [==============================] - 0s 779us/sample - loss: 0.5674 - accuracy: 0.7500\n",
      "Epoch 630/1000\n",
      "12/12 [==============================] - 0s 712us/sample - loss: 1.0947 - accuracy: 0.5000\n",
      "Epoch 631/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 0.5163 - accuracy: 0.8333\n",
      "Epoch 632/1000\n",
      "12/12 [==============================] - 0s 713us/sample - loss: 0.7743 - accuracy: 0.5000\n",
      "Epoch 633/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 0.8087 - accuracy: 0.6667\n",
      "Epoch 634/1000\n",
      "12/12 [==============================] - 0s 819us/sample - loss: 0.5763 - accuracy: 0.9167\n",
      "Epoch 635/1000\n",
      "12/12 [==============================] - 0s 780us/sample - loss: 0.9012 - accuracy: 0.7500\n",
      "Epoch 636/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.5882 - accuracy: 0.7500\n",
      "Epoch 637/1000\n",
      "12/12 [==============================] - 0s 735us/sample - loss: 0.5424 - accuracy: 0.7500\n",
      "Epoch 638/1000\n",
      "12/12 [==============================] - 0s 800us/sample - loss: 1.0651 - accuracy: 0.5000\n",
      "Epoch 639/1000\n",
      "12/12 [==============================] - 0s 821us/sample - loss: 0.6759 - accuracy: 0.7500\n",
      "Epoch 640/1000\n",
      "12/12 [==============================] - 0s 724us/sample - loss: 0.8272 - accuracy: 0.6667\n",
      "Epoch 641/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 1.3059 - accuracy: 0.5000\n",
      "Epoch 642/1000\n",
      "12/12 [==============================] - 0s 751us/sample - loss: 0.4473 - accuracy: 0.8333\n",
      "Epoch 643/1000\n",
      "12/12 [==============================] - 0s 736us/sample - loss: 0.8739 - accuracy: 0.5833\n",
      "Epoch 644/1000\n",
      "12/12 [==============================] - 0s 746us/sample - loss: 0.4694 - accuracy: 0.8333\n",
      "Epoch 645/1000\n",
      "12/12 [==============================] - 0s 795us/sample - loss: 0.7638 - accuracy: 0.7500\n",
      "Epoch 646/1000\n",
      "12/12 [==============================] - 0s 808us/sample - loss: 0.6824 - accuracy: 0.6667\n",
      "Epoch 647/1000\n",
      "12/12 [==============================] - 0s 672us/sample - loss: 0.6566 - accuracy: 0.8333\n",
      "Epoch 648/1000\n",
      "12/12 [==============================] - 0s 839us/sample - loss: 1.0405 - accuracy: 0.4167\n",
      "Epoch 649/1000\n",
      "12/12 [==============================] - 0s 799us/sample - loss: 0.9650 - accuracy: 0.6667\n",
      "Epoch 650/1000\n",
      "12/12 [==============================] - 0s 793us/sample - loss: 0.6339 - accuracy: 0.7500\n",
      "Epoch 651/1000\n",
      "12/12 [==============================] - 0s 753us/sample - loss: 0.6682 - accuracy: 0.6667\n",
      "Epoch 652/1000\n",
      "12/12 [==============================] - 0s 751us/sample - loss: 0.3146 - accuracy: 0.9167\n",
      "Epoch 653/1000\n",
      "12/12 [==============================] - 0s 753us/sample - loss: 0.6721 - accuracy: 0.5833\n",
      "Epoch 654/1000\n",
      "12/12 [==============================] - 0s 695us/sample - loss: 0.5335 - accuracy: 0.8333\n",
      "Epoch 655/1000\n",
      "12/12 [==============================] - 0s 775us/sample - loss: 0.5889 - accuracy: 0.9167\n",
      "Epoch 656/1000\n",
      "12/12 [==============================] - 0s 817us/sample - loss: 0.7729 - accuracy: 0.7500\n",
      "Epoch 657/1000\n",
      "12/12 [==============================] - 0s 708us/sample - loss: 0.4441 - accuracy: 0.8333\n",
      "Epoch 658/1000\n",
      "12/12 [==============================] - 0s 796us/sample - loss: 0.5223 - accuracy: 0.7500\n",
      "Epoch 659/1000\n",
      "12/12 [==============================] - 0s 828us/sample - loss: 0.4753 - accuracy: 0.8333\n",
      "Epoch 660/1000\n",
      "12/12 [==============================] - 0s 792us/sample - loss: 0.3404 - accuracy: 0.9167\n",
      "Epoch 661/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 0.8868 - accuracy: 0.6667\n",
      "Epoch 662/1000\n",
      "12/12 [==============================] - 0s 744us/sample - loss: 0.7516 - accuracy: 0.7500\n",
      "Epoch 663/1000\n",
      "12/12 [==============================] - 0s 736us/sample - loss: 0.7643 - accuracy: 0.6667\n",
      "Epoch 664/1000\n",
      "12/12 [==============================] - 0s 730us/sample - loss: 0.4843 - accuracy: 0.7500\n",
      "Epoch 665/1000\n",
      "12/12 [==============================] - 0s 765us/sample - loss: 0.4651 - accuracy: 0.8333\n",
      "Epoch 666/1000\n",
      "12/12 [==============================] - 0s 772us/sample - loss: 0.8715 - accuracy: 0.5833\n",
      "Epoch 667/1000\n",
      "12/12 [==============================] - 0s 719us/sample - loss: 0.6141 - accuracy: 0.7500\n",
      "Epoch 668/1000\n",
      "12/12 [==============================] - 0s 679us/sample - loss: 0.6816 - accuracy: 0.5833\n",
      "Epoch 669/1000\n",
      "12/12 [==============================] - 0s 723us/sample - loss: 0.3495 - accuracy: 0.8333\n",
      "Epoch 670/1000\n",
      "12/12 [==============================] - 0s 783us/sample - loss: 0.2596 - accuracy: 1.0000\n",
      "Epoch 671/1000\n",
      "12/12 [==============================] - 0s 694us/sample - loss: 0.6666 - accuracy: 0.7500\n",
      "Epoch 672/1000\n",
      "12/12 [==============================] - 0s 709us/sample - loss: 0.9012 - accuracy: 0.6667\n",
      "Epoch 673/1000\n",
      "12/12 [==============================] - 0s 806us/sample - loss: 0.3296 - accuracy: 1.0000\n",
      "Epoch 674/1000\n",
      "12/12 [==============================] - 0s 777us/sample - loss: 0.5502 - accuracy: 0.7500\n",
      "Epoch 675/1000\n",
      "12/12 [==============================] - 0s 716us/sample - loss: 0.5909 - accuracy: 0.8333\n",
      "Epoch 676/1000\n",
      "12/12 [==============================] - 0s 767us/sample - loss: 0.5740 - accuracy: 0.8333\n",
      "Epoch 677/1000\n",
      "12/12 [==============================] - 0s 770us/sample - loss: 0.4675 - accuracy: 0.8333\n",
      "Epoch 678/1000\n",
      "12/12 [==============================] - 0s 677us/sample - loss: 0.6193 - accuracy: 0.7500\n",
      "Epoch 679/1000\n",
      "12/12 [==============================] - 0s 693us/sample - loss: 0.9997 - accuracy: 0.5833\n",
      "Epoch 680/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 0.4742 - accuracy: 0.9167\n",
      "Epoch 681/1000\n",
      "12/12 [==============================] - 0s 811us/sample - loss: 0.7869 - accuracy: 0.4167\n",
      "Epoch 682/1000\n",
      "12/12 [==============================] - 0s 720us/sample - loss: 0.9248 - accuracy: 0.5000\n",
      "Epoch 683/1000\n",
      "12/12 [==============================] - 0s 800us/sample - loss: 1.3535 - accuracy: 0.2500\n",
      "Epoch 684/1000\n",
      "12/12 [==============================] - 0s 798us/sample - loss: 0.8403 - accuracy: 0.6667\n",
      "Epoch 685/1000\n",
      "12/12 [==============================] - 0s 738us/sample - loss: 0.7404 - accuracy: 0.6667\n",
      "Epoch 686/1000\n",
      "12/12 [==============================] - 0s 801us/sample - loss: 0.5940 - accuracy: 0.8333\n",
      "Epoch 687/1000\n",
      "12/12 [==============================] - 0s 792us/sample - loss: 0.7063 - accuracy: 0.5833\n",
      "Epoch 688/1000\n",
      "12/12 [==============================] - 0s 806us/sample - loss: 0.7640 - accuracy: 0.7500\n",
      "Epoch 689/1000\n",
      "12/12 [==============================] - 0s 847us/sample - loss: 0.5932 - accuracy: 0.8333\n",
      "Epoch 690/1000\n",
      "12/12 [==============================] - 0s 858us/sample - loss: 0.2098 - accuracy: 0.9167\n",
      "Epoch 691/1000\n",
      "12/12 [==============================] - 0s 831us/sample - loss: 0.3570 - accuracy: 0.8333\n",
      "Epoch 692/1000\n",
      "12/12 [==============================] - 0s 880us/sample - loss: 0.9938 - accuracy: 0.5833\n",
      "Epoch 693/1000\n",
      "12/12 [==============================] - 0s 910us/sample - loss: 0.6053 - accuracy: 0.7500\n",
      "Epoch 694/1000\n",
      "12/12 [==============================] - 0s 835us/sample - loss: 0.7631 - accuracy: 0.5000\n",
      "Epoch 695/1000\n",
      "12/12 [==============================] - 0s 881us/sample - loss: 0.5302 - accuracy: 0.7500\n",
      "Epoch 696/1000\n",
      "12/12 [==============================] - 0s 902us/sample - loss: 0.3891 - accuracy: 0.8333\n",
      "Epoch 697/1000\n",
      "12/12 [==============================] - 0s 866us/sample - loss: 0.7518 - accuracy: 0.7500\n",
      "Epoch 698/1000\n",
      "12/12 [==============================] - 0s 887us/sample - loss: 1.0816 - accuracy: 0.3333\n",
      "Epoch 699/1000\n",
      "12/12 [==============================] - 0s 933us/sample - loss: 0.8266 - accuracy: 0.5833\n",
      "Epoch 700/1000\n",
      "12/12 [==============================] - 0s 882us/sample - loss: 1.0011 - accuracy: 0.4167\n",
      "Epoch 701/1000\n",
      "12/12 [==============================] - 0s 912us/sample - loss: 0.6484 - accuracy: 0.7500\n",
      "Epoch 702/1000\n",
      "12/12 [==============================] - 0s 879us/sample - loss: 1.0945 - accuracy: 0.5833\n",
      "Epoch 703/1000\n",
      "12/12 [==============================] - 0s 874us/sample - loss: 0.3625 - accuracy: 0.9167\n",
      "Epoch 704/1000\n",
      "12/12 [==============================] - 0s 827us/sample - loss: 0.9116 - accuracy: 0.4167\n",
      "Epoch 705/1000\n",
      "12/12 [==============================] - 0s 797us/sample - loss: 1.0579 - accuracy: 0.6667\n",
      "Epoch 706/1000\n",
      "12/12 [==============================] - 0s 761us/sample - loss: 1.0553 - accuracy: 0.6667\n",
      "Epoch 707/1000\n",
      "12/12 [==============================] - 0s 727us/sample - loss: 0.5107 - accuracy: 0.6667\n",
      "Epoch 708/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 0.5408 - accuracy: 0.7500\n",
      "Epoch 709/1000\n",
      "12/12 [==============================] - 0s 796us/sample - loss: 1.0713 - accuracy: 0.5000\n",
      "Epoch 710/1000\n",
      "12/12 [==============================] - 0s 683us/sample - loss: 1.0712 - accuracy: 0.5833\n",
      "Epoch 711/1000\n",
      "12/12 [==============================] - 0s 727us/sample - loss: 0.8336 - accuracy: 0.6667\n",
      "Epoch 712/1000\n",
      "12/12 [==============================] - 0s 762us/sample - loss: 0.9650 - accuracy: 0.5000\n",
      "Epoch 713/1000\n",
      "12/12 [==============================] - 0s 765us/sample - loss: 0.5974 - accuracy: 0.7500\n",
      "Epoch 714/1000\n",
      "12/12 [==============================] - 0s 705us/sample - loss: 0.5770 - accuracy: 0.8333\n",
      "Epoch 715/1000\n",
      "12/12 [==============================] - 0s 814us/sample - loss: 1.2915 - accuracy: 0.3333\n",
      "Epoch 716/1000\n",
      "12/12 [==============================] - 0s 767us/sample - loss: 0.8774 - accuracy: 0.5833\n",
      "Epoch 717/1000\n",
      "12/12 [==============================] - 0s 708us/sample - loss: 0.3402 - accuracy: 0.9167\n",
      "Epoch 718/1000\n",
      "12/12 [==============================] - 0s 792us/sample - loss: 0.7238 - accuracy: 0.6667\n",
      "Epoch 719/1000\n",
      "12/12 [==============================] - 0s 903us/sample - loss: 0.8367 - accuracy: 0.5833\n",
      "Epoch 720/1000\n",
      "12/12 [==============================] - 0s 743us/sample - loss: 0.9587 - accuracy: 0.5000\n",
      "Epoch 721/1000\n",
      "12/12 [==============================] - 0s 723us/sample - loss: 0.8077 - accuracy: 0.7500\n",
      "Epoch 722/1000\n",
      "12/12 [==============================] - 0s 809us/sample - loss: 0.7582 - accuracy: 0.5833\n",
      "Epoch 723/1000\n",
      "12/12 [==============================] - 0s 773us/sample - loss: 0.7911 - accuracy: 0.6667\n",
      "Epoch 724/1000\n",
      "12/12 [==============================] - 0s 682us/sample - loss: 0.2575 - accuracy: 0.9167\n",
      "Epoch 725/1000\n",
      "12/12 [==============================] - 0s 849us/sample - loss: 0.9035 - accuracy: 0.5833\n",
      "Epoch 726/1000\n",
      "12/12 [==============================] - 0s 855us/sample - loss: 0.6093 - accuracy: 0.7500\n",
      "Epoch 727/1000\n",
      "12/12 [==============================] - 0s 960us/sample - loss: 1.1883 - accuracy: 0.5000\n",
      "Epoch 728/1000\n",
      "12/12 [==============================] - 0s 928us/sample - loss: 0.8661 - accuracy: 0.5833\n",
      "Epoch 729/1000\n",
      "12/12 [==============================] - 0s 882us/sample - loss: 0.6062 - accuracy: 0.6667\n",
      "Epoch 730/1000\n",
      "12/12 [==============================] - 0s 704us/sample - loss: 0.8186 - accuracy: 0.7500\n",
      "Epoch 731/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 0.4406 - accuracy: 0.9167\n",
      "Epoch 732/1000\n",
      "12/12 [==============================] - 0s 763us/sample - loss: 0.5908 - accuracy: 0.8333\n",
      "Epoch 733/1000\n",
      "12/12 [==============================] - 0s 724us/sample - loss: 0.8705 - accuracy: 0.5833\n",
      "Epoch 734/1000\n",
      "12/12 [==============================] - 0s 749us/sample - loss: 1.0763 - accuracy: 0.5833\n",
      "Epoch 735/1000\n",
      "12/12 [==============================] - 0s 716us/sample - loss: 0.4766 - accuracy: 0.7500\n",
      "Epoch 736/1000\n",
      "12/12 [==============================] - 0s 724us/sample - loss: 0.8959 - accuracy: 0.6667\n",
      "Epoch 737/1000\n",
      "12/12 [==============================] - 0s 669us/sample - loss: 0.5887 - accuracy: 0.7500\n",
      "Epoch 738/1000\n",
      "12/12 [==============================] - 0s 695us/sample - loss: 0.7360 - accuracy: 0.7500\n",
      "Epoch 739/1000\n",
      "12/12 [==============================] - 0s 755us/sample - loss: 1.1660 - accuracy: 0.5000\n",
      "Epoch 740/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 0.8430 - accuracy: 0.5833\n",
      "Epoch 741/1000\n",
      "12/12 [==============================] - 0s 659us/sample - loss: 0.6669 - accuracy: 0.5833\n",
      "Epoch 742/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 0.6322 - accuracy: 0.6667\n",
      "Epoch 743/1000\n",
      "12/12 [==============================] - 0s 780us/sample - loss: 0.3814 - accuracy: 0.9167\n",
      "Epoch 744/1000\n",
      "12/12 [==============================] - 0s 692us/sample - loss: 0.9037 - accuracy: 0.5833\n",
      "Epoch 745/1000\n",
      "12/12 [==============================] - 0s 733us/sample - loss: 0.6907 - accuracy: 0.6667\n",
      "Epoch 746/1000\n",
      "12/12 [==============================] - 0s 786us/sample - loss: 0.8344 - accuracy: 0.5833\n",
      "Epoch 747/1000\n",
      "12/12 [==============================] - 0s 782us/sample - loss: 1.0559 - accuracy: 0.5000\n",
      "Epoch 748/1000\n",
      "12/12 [==============================] - 0s 643us/sample - loss: 0.5243 - accuracy: 0.7500\n",
      "Epoch 749/1000\n",
      "12/12 [==============================] - 0s 871us/sample - loss: 0.7744 - accuracy: 0.5833\n",
      "Epoch 750/1000\n",
      "12/12 [==============================] - 0s 825us/sample - loss: 0.6831 - accuracy: 0.6667\n",
      "Epoch 751/1000\n",
      "12/12 [==============================] - 0s 744us/sample - loss: 0.8055 - accuracy: 0.5833\n",
      "Epoch 752/1000\n",
      "12/12 [==============================] - 0s 803us/sample - loss: 0.6870 - accuracy: 0.5833\n",
      "Epoch 753/1000\n",
      "12/12 [==============================] - 0s 714us/sample - loss: 0.4752 - accuracy: 0.8333\n",
      "Epoch 754/1000\n",
      "12/12 [==============================] - 0s 739us/sample - loss: 0.5877 - accuracy: 0.7500\n",
      "Epoch 755/1000\n",
      "12/12 [==============================] - 0s 693us/sample - loss: 0.7461 - accuracy: 0.5000\n",
      "Epoch 756/1000\n",
      "12/12 [==============================] - 0s 914us/sample - loss: 0.9832 - accuracy: 0.5000\n",
      "Epoch 757/1000\n",
      "12/12 [==============================] - 0s 794us/sample - loss: 0.6868 - accuracy: 0.8333\n",
      "Epoch 758/1000\n",
      "12/12 [==============================] - 0s 679us/sample - loss: 0.5649 - accuracy: 0.8333\n",
      "Epoch 759/1000\n",
      "12/12 [==============================] - 0s 773us/sample - loss: 0.8328 - accuracy: 0.6667\n",
      "Epoch 760/1000\n",
      "12/12 [==============================] - 0s 744us/sample - loss: 0.7620 - accuracy: 0.6667\n",
      "Epoch 761/1000\n",
      "12/12 [==============================] - 0s 766us/sample - loss: 0.8780 - accuracy: 0.5000\n",
      "Epoch 762/1000\n",
      "12/12 [==============================] - 0s 763us/sample - loss: 0.3351 - accuracy: 0.8333\n",
      "Epoch 763/1000\n",
      "12/12 [==============================] - 0s 847us/sample - loss: 0.7656 - accuracy: 0.5833\n",
      "Epoch 764/1000\n",
      "12/12 [==============================] - 0s 761us/sample - loss: 0.6755 - accuracy: 0.6667\n",
      "Epoch 765/1000\n",
      "12/12 [==============================] - 0s 730us/sample - loss: 0.6809 - accuracy: 0.6667\n",
      "Epoch 766/1000\n",
      "12/12 [==============================] - 0s 788us/sample - loss: 0.6001 - accuracy: 0.6667\n",
      "Epoch 767/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 0.6249 - accuracy: 0.6667\n",
      "Epoch 768/1000\n",
      "12/12 [==============================] - 0s 673us/sample - loss: 0.7484 - accuracy: 0.7500\n",
      "Epoch 769/1000\n",
      "12/12 [==============================] - 0s 760us/sample - loss: 0.7467 - accuracy: 0.8333\n",
      "Epoch 770/1000\n",
      "12/12 [==============================] - 0s 735us/sample - loss: 0.7845 - accuracy: 0.6667\n",
      "Epoch 771/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 0.7582 - accuracy: 0.6667\n",
      "Epoch 772/1000\n",
      "12/12 [==============================] - 0s 681us/sample - loss: 0.5388 - accuracy: 0.7500\n",
      "Epoch 773/1000\n",
      "12/12 [==============================] - 0s 788us/sample - loss: 0.3991 - accuracy: 0.8333\n",
      "Epoch 774/1000\n",
      "12/12 [==============================] - 0s 973us/sample - loss: 0.8514 - accuracy: 0.6667\n",
      "Epoch 775/1000\n",
      "12/12 [==============================] - 0s 806us/sample - loss: 0.7358 - accuracy: 0.7500\n",
      "Epoch 776/1000\n",
      "12/12 [==============================] - 0s 764us/sample - loss: 0.8985 - accuracy: 0.6667\n",
      "Epoch 777/1000\n",
      "12/12 [==============================] - 0s 826us/sample - loss: 0.5549 - accuracy: 0.5833\n",
      "Epoch 778/1000\n",
      "12/12 [==============================] - 0s 792us/sample - loss: 0.7103 - accuracy: 0.5000\n",
      "Epoch 779/1000\n",
      "12/12 [==============================] - 0s 775us/sample - loss: 0.7506 - accuracy: 0.5833\n",
      "Epoch 780/1000\n",
      "12/12 [==============================] - 0s 770us/sample - loss: 0.4915 - accuracy: 0.8333\n",
      "Epoch 781/1000\n",
      "12/12 [==============================] - 0s 842us/sample - loss: 0.9736 - accuracy: 0.6667\n",
      "Epoch 782/1000\n",
      "12/12 [==============================] - 0s 853us/sample - loss: 0.7892 - accuracy: 0.7500\n",
      "Epoch 783/1000\n",
      "12/12 [==============================] - 0s 727us/sample - loss: 0.6148 - accuracy: 0.7500\n",
      "Epoch 784/1000\n",
      "12/12 [==============================] - 0s 844us/sample - loss: 1.0415 - accuracy: 0.6667\n",
      "Epoch 785/1000\n",
      "12/12 [==============================] - 0s 774us/sample - loss: 0.9192 - accuracy: 0.5833\n",
      "Epoch 786/1000\n",
      "12/12 [==============================] - 0s 763us/sample - loss: 0.6254 - accuracy: 0.7500\n",
      "Epoch 787/1000\n",
      "12/12 [==============================] - 0s 723us/sample - loss: 0.9557 - accuracy: 0.5000\n",
      "Epoch 788/1000\n",
      "12/12 [==============================] - 0s 778us/sample - loss: 0.4283 - accuracy: 0.6667\n",
      "Epoch 789/1000\n",
      "12/12 [==============================] - 0s 908us/sample - loss: 0.6006 - accuracy: 0.6667\n",
      "Epoch 790/1000\n",
      "12/12 [==============================] - 0s 750us/sample - loss: 0.7877 - accuracy: 0.5833\n",
      "Epoch 791/1000\n",
      "12/12 [==============================] - 0s 883us/sample - loss: 0.5239 - accuracy: 0.8333\n",
      "Epoch 792/1000\n",
      "12/12 [==============================] - 0s 820us/sample - loss: 0.7811 - accuracy: 0.6667\n",
      "Epoch 793/1000\n",
      "12/12 [==============================] - 0s 697us/sample - loss: 1.0175 - accuracy: 0.5833\n",
      "Epoch 794/1000\n",
      "12/12 [==============================] - 0s 811us/sample - loss: 0.5697 - accuracy: 0.6667\n",
      "Epoch 795/1000\n",
      "12/12 [==============================] - 0s 790us/sample - loss: 0.5307 - accuracy: 0.7500\n",
      "Epoch 796/1000\n",
      "12/12 [==============================] - 0s 741us/sample - loss: 0.9386 - accuracy: 0.5000\n",
      "Epoch 797/1000\n",
      "12/12 [==============================] - 0s 749us/sample - loss: 0.3747 - accuracy: 0.9167\n",
      "Epoch 798/1000\n",
      "12/12 [==============================] - 0s 801us/sample - loss: 0.9046 - accuracy: 0.5833\n",
      "Epoch 799/1000\n",
      "12/12 [==============================] - 0s 774us/sample - loss: 0.5935 - accuracy: 0.7500\n",
      "Epoch 800/1000\n",
      "12/12 [==============================] - 0s 729us/sample - loss: 0.4409 - accuracy: 0.8333\n",
      "Epoch 801/1000\n",
      "12/12 [==============================] - 0s 745us/sample - loss: 0.6090 - accuracy: 0.6667\n",
      "Epoch 802/1000\n",
      "12/12 [==============================] - 0s 841us/sample - loss: 0.6272 - accuracy: 0.7500\n",
      "Epoch 803/1000\n",
      "12/12 [==============================] - 0s 668us/sample - loss: 0.7365 - accuracy: 0.5833\n",
      "Epoch 804/1000\n",
      "12/12 [==============================] - 0s 699us/sample - loss: 1.1091 - accuracy: 0.6667\n",
      "Epoch 805/1000\n",
      "12/12 [==============================] - 0s 761us/sample - loss: 0.7019 - accuracy: 0.5833\n",
      "Epoch 806/1000\n",
      "12/12 [==============================] - 0s 810us/sample - loss: 0.9406 - accuracy: 0.5000\n",
      "Epoch 807/1000\n",
      "12/12 [==============================] - 0s 719us/sample - loss: 0.3879 - accuracy: 0.7500\n",
      "Epoch 808/1000\n",
      "12/12 [==============================] - 0s 866us/sample - loss: 0.7086 - accuracy: 0.6667\n",
      "Epoch 809/1000\n",
      "12/12 [==============================] - 0s 854us/sample - loss: 0.8460 - accuracy: 0.5833\n",
      "Epoch 810/1000\n",
      "12/12 [==============================] - 0s 669us/sample - loss: 0.6171 - accuracy: 0.6667\n",
      "Epoch 811/1000\n",
      "12/12 [==============================] - 0s 710us/sample - loss: 0.6590 - accuracy: 0.5833\n",
      "Epoch 812/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.9761 - accuracy: 0.5000\n",
      "Epoch 813/1000\n",
      "12/12 [==============================] - 0s 723us/sample - loss: 0.7022 - accuracy: 0.6667\n",
      "Epoch 814/1000\n",
      "12/12 [==============================] - 0s 749us/sample - loss: 0.3043 - accuracy: 0.8333\n",
      "Epoch 815/1000\n",
      "12/12 [==============================] - 0s 807us/sample - loss: 1.1195 - accuracy: 0.4167\n",
      "Epoch 816/1000\n",
      "12/12 [==============================] - 0s 802us/sample - loss: 0.4924 - accuracy: 0.8333\n",
      "Epoch 817/1000\n",
      "12/12 [==============================] - 0s 680us/sample - loss: 0.8174 - accuracy: 0.5833\n",
      "Epoch 818/1000\n",
      "12/12 [==============================] - 0s 768us/sample - loss: 0.6495 - accuracy: 0.6667\n",
      "Epoch 819/1000\n",
      "12/12 [==============================] - 0s 845us/sample - loss: 0.6221 - accuracy: 0.8333\n",
      "Epoch 820/1000\n",
      "12/12 [==============================] - 0s 761us/sample - loss: 0.5937 - accuracy: 0.6667\n",
      "Epoch 821/1000\n",
      "12/12 [==============================] - 0s 756us/sample - loss: 0.5432 - accuracy: 0.7500\n",
      "Epoch 822/1000\n",
      "12/12 [==============================] - 0s 924us/sample - loss: 0.6429 - accuracy: 0.7500\n",
      "Epoch 823/1000\n",
      "12/12 [==============================] - 0s 740us/sample - loss: 0.5772 - accuracy: 0.8333\n",
      "Epoch 824/1000\n",
      "12/12 [==============================] - 0s 795us/sample - loss: 0.6387 - accuracy: 0.7500\n",
      "Epoch 825/1000\n",
      "12/12 [==============================] - 0s 799us/sample - loss: 0.9259 - accuracy: 0.6667\n",
      "Epoch 826/1000\n",
      "12/12 [==============================] - 0s 796us/sample - loss: 0.4625 - accuracy: 0.8333\n",
      "Epoch 827/1000\n",
      "12/12 [==============================] - 0s 678us/sample - loss: 0.7127 - accuracy: 0.6667\n",
      "Epoch 828/1000\n",
      "12/12 [==============================] - 0s 847us/sample - loss: 0.3146 - accuracy: 0.8333\n",
      "Epoch 829/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.7486 - accuracy: 0.5000\n",
      "Epoch 830/1000\n",
      "12/12 [==============================] - 0s 925us/sample - loss: 0.5584 - accuracy: 0.7500\n",
      "Epoch 831/1000\n",
      "12/12 [==============================] - 0s 816us/sample - loss: 0.5304 - accuracy: 0.7500\n",
      "Epoch 832/1000\n",
      "12/12 [==============================] - 0s 834us/sample - loss: 0.3836 - accuracy: 0.8333\n",
      "Epoch 833/1000\n",
      "12/12 [==============================] - 0s 777us/sample - loss: 0.5899 - accuracy: 0.6667\n",
      "Epoch 834/1000\n",
      "12/12 [==============================] - 0s 716us/sample - loss: 0.8439 - accuracy: 0.5000\n",
      "Epoch 835/1000\n",
      "12/12 [==============================] - 0s 806us/sample - loss: 0.5102 - accuracy: 0.7500\n",
      "Epoch 836/1000\n",
      "12/12 [==============================] - 0s 790us/sample - loss: 0.6507 - accuracy: 0.7500\n",
      "Epoch 837/1000\n",
      "12/12 [==============================] - 0s 651us/sample - loss: 0.3529 - accuracy: 0.9167\n",
      "Epoch 838/1000\n",
      "12/12 [==============================] - 0s 772us/sample - loss: 0.3580 - accuracy: 0.8333\n",
      "Epoch 839/1000\n",
      "12/12 [==============================] - 0s 756us/sample - loss: 0.9123 - accuracy: 0.4167\n",
      "Epoch 840/1000\n",
      "12/12 [==============================] - 0s 724us/sample - loss: 0.8015 - accuracy: 0.5833\n",
      "Epoch 841/1000\n",
      "12/12 [==============================] - 0s 691us/sample - loss: 0.6163 - accuracy: 0.6667\n",
      "Epoch 842/1000\n",
      "12/12 [==============================] - 0s 724us/sample - loss: 0.9604 - accuracy: 0.5833\n",
      "Epoch 843/1000\n",
      "12/12 [==============================] - 0s 731us/sample - loss: 0.4044 - accuracy: 0.8333\n",
      "Epoch 844/1000\n",
      "12/12 [==============================] - 0s 654us/sample - loss: 0.7118 - accuracy: 0.6667\n",
      "Epoch 845/1000\n",
      "12/12 [==============================] - 0s 724us/sample - loss: 0.8641 - accuracy: 0.8333\n",
      "Epoch 846/1000\n",
      "12/12 [==============================] - 0s 723us/sample - loss: 0.5889 - accuracy: 0.6667\n",
      "Epoch 847/1000\n",
      "12/12 [==============================] - 0s 712us/sample - loss: 0.6604 - accuracy: 0.6667\n",
      "Epoch 848/1000\n",
      "12/12 [==============================] - 0s 692us/sample - loss: 0.7002 - accuracy: 0.6667\n",
      "Epoch 849/1000\n",
      "12/12 [==============================] - 0s 738us/sample - loss: 0.5463 - accuracy: 0.7500\n",
      "Epoch 850/1000\n",
      "12/12 [==============================] - 0s 768us/sample - loss: 0.5246 - accuracy: 0.8333\n",
      "Epoch 851/1000\n",
      "12/12 [==============================] - 0s 667us/sample - loss: 0.6157 - accuracy: 0.6667\n",
      "Epoch 852/1000\n",
      "12/12 [==============================] - 0s 710us/sample - loss: 0.6207 - accuracy: 0.7500\n",
      "Epoch 853/1000\n",
      "12/12 [==============================] - 0s 727us/sample - loss: 0.6381 - accuracy: 0.7500\n",
      "Epoch 854/1000\n",
      "12/12 [==============================] - 0s 792us/sample - loss: 0.8038 - accuracy: 0.5833\n",
      "Epoch 855/1000\n",
      "12/12 [==============================] - 0s 661us/sample - loss: 0.7893 - accuracy: 0.5833\n",
      "Epoch 856/1000\n",
      "12/12 [==============================] - 0s 717us/sample - loss: 0.8684 - accuracy: 0.6667\n",
      "Epoch 857/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 1.1568 - accuracy: 0.5000\n",
      "Epoch 858/1000\n",
      "12/12 [==============================] - 0s 702us/sample - loss: 0.6147 - accuracy: 0.6667\n",
      "Epoch 859/1000\n",
      "12/12 [==============================] - 0s 687us/sample - loss: 0.4451 - accuracy: 0.8333\n",
      "Epoch 860/1000\n",
      "12/12 [==============================] - 0s 721us/sample - loss: 0.8810 - accuracy: 0.5833\n",
      "Epoch 861/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 0.3944 - accuracy: 0.8333\n",
      "Epoch 862/1000\n",
      "12/12 [==============================] - 0s 699us/sample - loss: 0.7770 - accuracy: 0.7500\n",
      "Epoch 863/1000\n",
      "12/12 [==============================] - 0s 674us/sample - loss: 0.8397 - accuracy: 0.6667\n",
      "Epoch 864/1000\n",
      "12/12 [==============================] - 0s 759us/sample - loss: 0.6902 - accuracy: 0.6667\n",
      "Epoch 865/1000\n",
      "12/12 [==============================] - 0s 737us/sample - loss: 1.0516 - accuracy: 0.5000\n",
      "Epoch 866/1000\n",
      "12/12 [==============================] - 0s 661us/sample - loss: 0.7005 - accuracy: 0.6667\n",
      "Epoch 867/1000\n",
      "12/12 [==============================] - 0s 736us/sample - loss: 0.5528 - accuracy: 0.7500\n",
      "Epoch 868/1000\n",
      "12/12 [==============================] - 0s 738us/sample - loss: 0.8207 - accuracy: 0.5833\n",
      "Epoch 869/1000\n",
      "12/12 [==============================] - 0s 698us/sample - loss: 0.6819 - accuracy: 0.6667\n",
      "Epoch 870/1000\n",
      "12/12 [==============================] - 0s 698us/sample - loss: 1.1228 - accuracy: 0.5833\n",
      "Epoch 871/1000\n",
      "12/12 [==============================] - 0s 717us/sample - loss: 0.4469 - accuracy: 0.8333\n",
      "Epoch 872/1000\n",
      "12/12 [==============================] - 0s 832us/sample - loss: 0.6300 - accuracy: 0.7500\n",
      "Epoch 873/1000\n",
      "12/12 [==============================] - 0s 697us/sample - loss: 0.9816 - accuracy: 0.6667\n",
      "Epoch 874/1000\n",
      "12/12 [==============================] - 0s 670us/sample - loss: 1.1568 - accuracy: 0.5000\n",
      "Epoch 875/1000\n",
      "12/12 [==============================] - 0s 775us/sample - loss: 0.4210 - accuracy: 0.9167\n",
      "Epoch 876/1000\n",
      "12/12 [==============================] - 0s 775us/sample - loss: 0.2956 - accuracy: 1.0000\n",
      "Epoch 877/1000\n",
      "12/12 [==============================] - 0s 658us/sample - loss: 0.3439 - accuracy: 0.7500\n",
      "Epoch 878/1000\n",
      "12/12 [==============================] - 0s 751us/sample - loss: 0.6374 - accuracy: 0.8333\n",
      "Epoch 879/1000\n",
      "12/12 [==============================] - 0s 753us/sample - loss: 0.9091 - accuracy: 0.6667\n",
      "Epoch 880/1000\n",
      "12/12 [==============================] - 0s 698us/sample - loss: 0.8979 - accuracy: 0.6667\n",
      "Epoch 881/1000\n",
      "12/12 [==============================] - 0s 680us/sample - loss: 0.8260 - accuracy: 0.6667\n",
      "Epoch 882/1000\n",
      "12/12 [==============================] - 0s 757us/sample - loss: 0.6396 - accuracy: 0.6667\n",
      "Epoch 883/1000\n",
      "12/12 [==============================] - 0s 747us/sample - loss: 0.4339 - accuracy: 0.8333\n",
      "Epoch 884/1000\n",
      "12/12 [==============================] - 0s 670us/sample - loss: 0.7447 - accuracy: 0.6667\n",
      "Epoch 885/1000\n",
      "12/12 [==============================] - 0s 759us/sample - loss: 0.5419 - accuracy: 0.7500\n",
      "Epoch 886/1000\n",
      "12/12 [==============================] - 0s 716us/sample - loss: 0.5431 - accuracy: 0.7500\n",
      "Epoch 887/1000\n",
      "12/12 [==============================] - 0s 702us/sample - loss: 0.6524 - accuracy: 0.6667\n",
      "Epoch 888/1000\n",
      "12/12 [==============================] - 0s 680us/sample - loss: 0.7800 - accuracy: 0.5833\n",
      "Epoch 889/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.9450 - accuracy: 0.5000\n",
      "Epoch 890/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 0.4693 - accuracy: 0.9167\n",
      "Epoch 891/1000\n",
      "12/12 [==============================] - 0s 656us/sample - loss: 0.9458 - accuracy: 0.5000\n",
      "Epoch 892/1000\n",
      "12/12 [==============================] - 0s 678us/sample - loss: 0.6430 - accuracy: 0.6667\n",
      "Epoch 893/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.4342 - accuracy: 0.8333\n",
      "Epoch 894/1000\n",
      "12/12 [==============================] - 0s 765us/sample - loss: 0.4674 - accuracy: 0.8333\n",
      "Epoch 895/1000\n",
      "12/12 [==============================] - 0s 728us/sample - loss: 1.0376 - accuracy: 0.4167\n",
      "Epoch 896/1000\n",
      "12/12 [==============================] - 0s 783us/sample - loss: 0.6701 - accuracy: 0.7500\n",
      "Epoch 897/1000\n",
      "12/12 [==============================] - 0s 716us/sample - loss: 0.6517 - accuracy: 0.8333\n",
      "Epoch 898/1000\n",
      "12/12 [==============================] - 0s 667us/sample - loss: 0.6211 - accuracy: 0.8333\n",
      "Epoch 899/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.4600 - accuracy: 0.8333\n",
      "Epoch 900/1000\n",
      "12/12 [==============================] - 0s 796us/sample - loss: 0.6442 - accuracy: 0.8333\n",
      "Epoch 901/1000\n",
      "12/12 [==============================] - 0s 772us/sample - loss: 0.3677 - accuracy: 0.8333\n",
      "Epoch 902/1000\n",
      "12/12 [==============================] - 0s 645us/sample - loss: 0.9052 - accuracy: 0.5000\n",
      "Epoch 903/1000\n",
      "12/12 [==============================] - 0s 747us/sample - loss: 0.5551 - accuracy: 0.7500\n",
      "Epoch 904/1000\n",
      "12/12 [==============================] - 0s 716us/sample - loss: 0.9099 - accuracy: 0.5000\n",
      "Epoch 905/1000\n",
      "12/12 [==============================] - 0s 724us/sample - loss: 0.5022 - accuracy: 0.8333\n",
      "Epoch 906/1000\n",
      "12/12 [==============================] - 0s 683us/sample - loss: 0.7821 - accuracy: 0.6667\n",
      "Epoch 907/1000\n",
      "12/12 [==============================] - 0s 767us/sample - loss: 0.5311 - accuracy: 0.8333\n",
      "Epoch 908/1000\n",
      "12/12 [==============================] - 0s 812us/sample - loss: 0.3801 - accuracy: 0.8333\n",
      "Epoch 909/1000\n",
      "12/12 [==============================] - 0s 697us/sample - loss: 0.3980 - accuracy: 0.8333\n",
      "Epoch 910/1000\n",
      "12/12 [==============================] - 0s 730us/sample - loss: 0.7163 - accuracy: 0.6667\n",
      "Epoch 911/1000\n",
      "12/12 [==============================] - 0s 749us/sample - loss: 1.2780 - accuracy: 0.5000\n",
      "Epoch 912/1000\n",
      "12/12 [==============================] - 0s 756us/sample - loss: 1.0350 - accuracy: 0.5000\n",
      "Epoch 913/1000\n",
      "12/12 [==============================] - 0s 655us/sample - loss: 0.6534 - accuracy: 0.7500\n",
      "Epoch 914/1000\n",
      "12/12 [==============================] - 0s 726us/sample - loss: 0.6675 - accuracy: 0.6667\n",
      "Epoch 915/1000\n",
      "12/12 [==============================] - 0s 736us/sample - loss: 0.7275 - accuracy: 0.6667\n",
      "Epoch 916/1000\n",
      "12/12 [==============================] - 0s 677us/sample - loss: 0.8076 - accuracy: 0.6667\n",
      "Epoch 917/1000\n",
      "12/12 [==============================] - 0s 727us/sample - loss: 0.6666 - accuracy: 0.7500\n",
      "Epoch 918/1000\n",
      "12/12 [==============================] - 0s 757us/sample - loss: 0.9074 - accuracy: 0.6667\n",
      "Epoch 919/1000\n",
      "12/12 [==============================] - 0s 723us/sample - loss: 0.9129 - accuracy: 0.5833\n",
      "Epoch 920/1000\n",
      "12/12 [==============================] - 0s 669us/sample - loss: 0.3259 - accuracy: 0.9167\n",
      "Epoch 921/1000\n",
      "12/12 [==============================] - 0s 697us/sample - loss: 0.5206 - accuracy: 0.7500\n",
      "Epoch 922/1000\n",
      "12/12 [==============================] - 0s 746us/sample - loss: 0.7992 - accuracy: 0.5833\n",
      "Epoch 923/1000\n",
      "12/12 [==============================] - 0s 702us/sample - loss: 0.5864 - accuracy: 0.7500\n",
      "Epoch 924/1000\n",
      "12/12 [==============================] - 0s 692us/sample - loss: 0.5780 - accuracy: 0.7500\n",
      "Epoch 925/1000\n",
      "12/12 [==============================] - 0s 777us/sample - loss: 0.5973 - accuracy: 0.5833\n",
      "Epoch 926/1000\n",
      "12/12 [==============================] - 0s 780us/sample - loss: 0.7647 - accuracy: 0.7500\n",
      "Epoch 927/1000\n",
      "12/12 [==============================] - 0s 664us/sample - loss: 0.3952 - accuracy: 0.9167\n",
      "Epoch 928/1000\n",
      "12/12 [==============================] - 0s 691us/sample - loss: 0.8504 - accuracy: 0.6667\n",
      "Epoch 929/1000\n",
      "12/12 [==============================] - 0s 770us/sample - loss: 0.7693 - accuracy: 0.6667\n",
      "Epoch 930/1000\n",
      "12/12 [==============================] - 0s 758us/sample - loss: 0.7135 - accuracy: 0.5833\n",
      "Epoch 931/1000\n",
      "12/12 [==============================] - 0s 727us/sample - loss: 0.6408 - accuracy: 0.6667\n",
      "Epoch 932/1000\n",
      "12/12 [==============================] - 0s 730us/sample - loss: 0.9006 - accuracy: 0.6667\n",
      "Epoch 933/1000\n",
      "12/12 [==============================] - 0s 782us/sample - loss: 1.0701 - accuracy: 0.5000\n",
      "Epoch 934/1000\n",
      "12/12 [==============================] - 0s 648us/sample - loss: 0.6323 - accuracy: 0.6667\n",
      "Epoch 935/1000\n",
      "12/12 [==============================] - 0s 666us/sample - loss: 0.5745 - accuracy: 0.6667\n",
      "Epoch 936/1000\n",
      "12/12 [==============================] - 0s 752us/sample - loss: 0.7836 - accuracy: 0.5000\n",
      "Epoch 937/1000\n",
      "12/12 [==============================] - 0s 737us/sample - loss: 0.9553 - accuracy: 0.5833\n",
      "Epoch 938/1000\n",
      "12/12 [==============================] - 0s 664us/sample - loss: 0.8876 - accuracy: 0.6667\n",
      "Epoch 939/1000\n",
      "12/12 [==============================] - 0s 698us/sample - loss: 1.3559 - accuracy: 0.3333\n",
      "Epoch 940/1000\n",
      "12/12 [==============================] - 0s 761us/sample - loss: 0.6658 - accuracy: 0.7500\n",
      "Epoch 941/1000\n",
      "12/12 [==============================] - 0s 741us/sample - loss: 0.9279 - accuracy: 0.5000\n",
      "Epoch 942/1000\n",
      "12/12 [==============================] - 0s 671us/sample - loss: 0.8416 - accuracy: 0.5833\n",
      "Epoch 943/1000\n",
      "12/12 [==============================] - 0s 747us/sample - loss: 0.6184 - accuracy: 0.6667\n",
      "Epoch 944/1000\n",
      "12/12 [==============================] - 0s 768us/sample - loss: 0.3572 - accuracy: 0.8333\n",
      "Epoch 945/1000\n",
      "12/12 [==============================] - 0s 650us/sample - loss: 1.0483 - accuracy: 0.4167\n",
      "Epoch 946/1000\n",
      "12/12 [==============================] - 0s 737us/sample - loss: 0.5948 - accuracy: 0.8333\n",
      "Epoch 947/1000\n",
      "12/12 [==============================] - 0s 798us/sample - loss: 0.7085 - accuracy: 0.5833\n",
      "Epoch 948/1000\n",
      "12/12 [==============================] - 0s 769us/sample - loss: 0.3158 - accuracy: 0.8333\n",
      "Epoch 949/1000\n",
      "12/12 [==============================] - 0s 678us/sample - loss: 0.8816 - accuracy: 0.6667\n",
      "Epoch 950/1000\n",
      "12/12 [==============================] - 0s 790us/sample - loss: 0.6484 - accuracy: 0.7500\n",
      "Epoch 951/1000\n",
      "12/12 [==============================] - 0s 800us/sample - loss: 0.9270 - accuracy: 0.5833\n",
      "Epoch 952/1000\n",
      "12/12 [==============================] - 0s 669us/sample - loss: 1.0492 - accuracy: 0.5000\n",
      "Epoch 953/1000\n",
      "12/12 [==============================] - 0s 749us/sample - loss: 0.4630 - accuracy: 0.8333\n",
      "Epoch 954/1000\n",
      "12/12 [==============================] - 0s 740us/sample - loss: 1.0766 - accuracy: 0.4167\n",
      "Epoch 955/1000\n",
      "12/12 [==============================] - 0s 776us/sample - loss: 0.4868 - accuracy: 0.7500\n",
      "Epoch 956/1000\n",
      "12/12 [==============================] - 0s 682us/sample - loss: 0.7284 - accuracy: 0.7500\n",
      "Epoch 957/1000\n",
      "12/12 [==============================] - 0s 745us/sample - loss: 0.6083 - accuracy: 0.7500\n",
      "Epoch 958/1000\n",
      "12/12 [==============================] - 0s 778us/sample - loss: 0.5978 - accuracy: 0.7500\n",
      "Epoch 959/1000\n",
      "12/12 [==============================] - 0s 730us/sample - loss: 1.2314 - accuracy: 0.4167\n",
      "Epoch 960/1000\n",
      "12/12 [==============================] - 0s 772us/sample - loss: 0.3772 - accuracy: 0.8333\n",
      "Epoch 961/1000\n",
      "12/12 [==============================] - 0s 805us/sample - loss: 0.5593 - accuracy: 0.6667\n",
      "Epoch 962/1000\n",
      "12/12 [==============================] - 0s 796us/sample - loss: 0.6418 - accuracy: 0.7500\n",
      "Epoch 963/1000\n",
      "12/12 [==============================] - 0s 705us/sample - loss: 0.7745 - accuracy: 0.6667\n",
      "Epoch 964/1000\n",
      "12/12 [==============================] - 0s 696us/sample - loss: 0.4776 - accuracy: 0.7500\n",
      "Epoch 965/1000\n",
      "12/12 [==============================] - 0s 813us/sample - loss: 0.5438 - accuracy: 0.8333\n",
      "Epoch 966/1000\n",
      "12/12 [==============================] - 0s 731us/sample - loss: 0.5104 - accuracy: 0.7500\n",
      "Epoch 967/1000\n",
      "12/12 [==============================] - 0s 700us/sample - loss: 0.6514 - accuracy: 0.5833\n",
      "Epoch 968/1000\n",
      "12/12 [==============================] - 0s 738us/sample - loss: 0.6224 - accuracy: 0.7500\n",
      "Epoch 969/1000\n",
      "12/12 [==============================] - 0s 779us/sample - loss: 1.1307 - accuracy: 0.4167\n",
      "Epoch 970/1000\n",
      "12/12 [==============================] - ETA: 0s - loss: 1.6977 - accuracy: 0.0000e+ - 0s 689us/sample - loss: 0.9067 - accuracy: 0.5000\n",
      "Epoch 971/1000\n",
      "12/12 [==============================] - 0s 731us/sample - loss: 0.1026 - accuracy: 1.0000\n",
      "Epoch 972/1000\n",
      "12/12 [==============================] - 0s 777us/sample - loss: 0.7486 - accuracy: 0.5833\n",
      "Epoch 973/1000\n",
      "12/12 [==============================] - 0s 719us/sample - loss: 0.6375 - accuracy: 0.6667\n",
      "Epoch 974/1000\n",
      "12/12 [==============================] - 0s 742us/sample - loss: 0.5017 - accuracy: 0.6667\n",
      "Epoch 975/1000\n",
      "12/12 [==============================] - 0s 766us/sample - loss: 0.8739 - accuracy: 0.4167\n",
      "Epoch 976/1000\n",
      "12/12 [==============================] - 0s 737us/sample - loss: 0.4932 - accuracy: 0.7500\n",
      "Epoch 977/1000\n",
      "12/12 [==============================] - 0s 665us/sample - loss: 0.5687 - accuracy: 0.5833\n",
      "Epoch 978/1000\n",
      "12/12 [==============================] - 0s 768us/sample - loss: 1.2891 - accuracy: 0.3333\n",
      "Epoch 979/1000\n",
      "12/12 [==============================] - 0s 782us/sample - loss: 0.8716 - accuracy: 0.5833\n",
      "Epoch 980/1000\n",
      "12/12 [==============================] - 0s 700us/sample - loss: 0.3516 - accuracy: 0.8333\n",
      "Epoch 981/1000\n",
      "12/12 [==============================] - 0s 844us/sample - loss: 0.4668 - accuracy: 0.7500\n",
      "Epoch 982/1000\n",
      "12/12 [==============================] - 0s 904us/sample - loss: 0.7487 - accuracy: 0.5833\n",
      "Epoch 983/1000\n",
      "12/12 [==============================] - 0s 895us/sample - loss: 0.5866 - accuracy: 0.7500\n",
      "Epoch 984/1000\n",
      "12/12 [==============================] - 0s 883us/sample - loss: 0.4271 - accuracy: 0.7500\n",
      "Epoch 985/1000\n",
      "12/12 [==============================] - 0s 940us/sample - loss: 0.7085 - accuracy: 0.6667\n",
      "Epoch 986/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 0.8176 - accuracy: 0.5000\n",
      "Epoch 987/1000\n",
      "12/12 [==============================] - 0s 914us/sample - loss: 0.5576 - accuracy: 0.8333\n",
      "Epoch 988/1000\n",
      "12/12 [==============================] - 0s 1ms/sample - loss: 0.8911 - accuracy: 0.6667\n",
      "Epoch 989/1000\n",
      "12/12 [==============================] - 0s 870us/sample - loss: 0.6170 - accuracy: 0.7500\n",
      "Epoch 990/1000\n",
      "12/12 [==============================] - 0s 924us/sample - loss: 0.4515 - accuracy: 0.9167\n",
      "Epoch 991/1000\n",
      "12/12 [==============================] - 0s 969us/sample - loss: 0.4017 - accuracy: 0.8333\n",
      "Epoch 992/1000\n",
      "12/12 [==============================] - 0s 830us/sample - loss: 0.4369 - accuracy: 0.8333\n",
      "Epoch 993/1000\n",
      "12/12 [==============================] - 0s 899us/sample - loss: 0.5738 - accuracy: 0.7500\n",
      "Epoch 994/1000\n",
      "12/12 [==============================] - 0s 854us/sample - loss: 0.7796 - accuracy: 0.5000\n",
      "Epoch 995/1000\n",
      "12/12 [==============================] - 0s 824us/sample - loss: 0.3730 - accuracy: 1.0000\n",
      "Epoch 996/1000\n",
      "12/12 [==============================] - 0s 942us/sample - loss: 0.8775 - accuracy: 0.5833\n",
      "Epoch 997/1000\n",
      "12/12 [==============================] - 0s 882us/sample - loss: 0.9376 - accuracy: 0.6667\n",
      "Epoch 998/1000\n",
      "12/12 [==============================] - 0s 857us/sample - loss: 0.7158 - accuracy: 0.7500\n",
      "Epoch 999/1000\n",
      "12/12 [==============================] - 0s 886us/sample - loss: 0.5577 - accuracy: 0.7500\n",
      "Epoch 1000/1000\n",
      "12/12 [==============================] - 0s 876us/sample - loss: 0.8531 - accuracy: 0.6667\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7f79b41d4e90>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train the model, iterating on the data in batches of 32 samples\n",
    "model.fit(train_x, train_y, nb_epoch=1000, batch_size=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'x_test' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-36-dd7bea78f917>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mclasses\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m128\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'x_test' is not defined"
     ]
    }
   ],
   "source": [
    "classes = model.predict(x_test, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}