prediction_Window_9_LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import keras\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.layers.core import Dense, Activation, Flatten, Dropout\n",
    "from keras.layers import Conv2D, LSTM, Embedding, Bidirectional\n",
    "from keras.models import Sequential\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import preprocessing\n",
    "from keras.utils import to_categorical\n",
    "from keras import regularizers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "ename": "IOError",
     "evalue": "[Errno 2] No such file or directory: 'fixed_sequences_length_9_PELM/Group_Phos_Y_neg.fasta'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mIOError\u001b[0m                                   Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-8522b0890db4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'fixed_sequences_length_9_PELM/Group_Phos_Y_neg.fasta'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'r'\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m     \u001b[0ms_negatif_txt\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreadlines\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0ms_negatif\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms_negatif_txt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mIOError\u001b[0m: [Errno 2] No such file or directory: 'fixed_sequences_length_9_PELM/Group_Phos_Y_neg.fasta'"
     ]
    }
   ],
   "source": [
    "with open('fixed_sequences_length_9_PELM/Group_Phos_Y_neg.fasta', 'r') as f:\n",
    "    s_negatif_txt = f.readlines()\n",
    "\n",
    "s_negatif = np.array([])\n",
    "for i in range(1,len(s_negatif_txt),2):\n",
    "    temp = s_negatif_txt[i]\n",
    "    temp1 = temp[0:9]\n",
    "    temp2 = list(temp1)\n",
    "    s_negatif = np.append(s_negatif, temp2)\n",
    "\n",
    "sequenceL = int(len(s_negatif)/9)\n",
    "s_negatif_new = np.expand_dims(s_negatif, axis=0)\n",
    "s_negatif_new = s_negatif_new.reshape(sequenceL,9)\n",
    "labelneg = np.zeros((sequenceL,), dtype=int)\n",
    "s_negatif_label = np.expand_dims(labelneg, axis=0)\n",
    "s_negatif_label = s_negatif_label.reshape(sequenceL,1)\n",
    "print(s_negatif_new.shape)\n",
    "print(s_negatif_label.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(267, 9)\n",
      "(267, 1)\n"
     ]
    }
   ],
   "source": [
    "with open('fixed_sequences_length_9_PELM/Group_Phos_Y_pos.fasta', 'r') as f:\n",
    "    s_positif_txt = f.readlines()\n",
    "\n",
    "s_positif = np.array([])\n",
    "for i in range(1,len(s_positif_txt),2):\n",
    "    temp = s_positif_txt[i]\n",
    "    temp1 = temp[0:9]\n",
    "    temp2 = list(temp1)\n",
    "    s_positif = np.append(s_positif, temp2)\n",
    "\n",
    "sequenceL = int(len(s_positif)/9)\n",
    "s_positif_new = np.expand_dims(s_positif, axis=0)\n",
    "s_positif_new = s_positif_new.reshape(sequenceL,9)\n",
    "labelpos = np.ones((sequenceL,), dtype=int)\n",
    "s_positif_label = np.expand_dims(labelpos, axis=0)\n",
    "s_positif_label = s_positif_label.reshape(sequenceL,1)\n",
    "print(s_positif_new.shape)\n",
    "print(s_positif_label.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(493, 9, 20, 1)\n",
      "(394, 9)\n",
      "[ 5.  1.  3. 10. 18. 15.  7. 10.  2.]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/bharuno/envtensorflow/lib/python3.5/site-packages/ipykernel_launcher.py:16: DeprecationWarning: Both axis > a.ndim and axis < -a.ndim - 1 are deprecated and will raise an AxisError in the future.\n",
      "  app.launch_new_instance()\n"
     ]
    }
   ],
   "source": [
    "training_X = np.concatenate((s_positif_new, s_negatif_new), axis=0, out=None)\n",
    "training_Y = np.concatenate((s_positif_label, s_negatif_label), axis=0, out=None)\n",
    "\n",
    "training_Y_onehot = to_categorical(training_Y)\n",
    "\n",
    "asam = ['A','R','N','D','C','Q','E','G','H','I','L','K','M','F','P','S','T','W','Y','V']\n",
    "tokenizer = Tokenizer(char_level=True)\n",
    "tokenizer.fit_on_texts(asam)\n",
    "sequence_of_int = []\n",
    "for i in range(len(training_X)):\n",
    "    temp = tokenizer.texts_to_sequences(training_X[i])\n",
    "    sequence_of_int = np.append(sequence_of_int, temp)\n",
    "\n",
    "sequence_of_int = sequence_of_int-1\n",
    "sequence_of_int_onehot = to_categorical(sequence_of_int)\n",
    "sequence_of_int_onehot = np.expand_dims(sequence_of_int_onehot, axis=3)\n",
    "sequence_of_int_onehot = sequence_of_int_onehot.reshape(len(training_X),9,20,1)\n",
    "sequence_of_int = sequence_of_int.reshape(len(training_X),9)\n",
    "\n",
    "print(sequence_of_int_onehot.shape)\n",
    "train_X, valid_X, train_label, valid_label = train_test_split(sequence_of_int, training_Y_onehot, test_size=0.2, random_state=13)\n",
    "print(train_X.shape)\n",
    "print(valid_X[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_17 (Embedding)     (None, 9, 8)              160       \n",
      "_________________________________________________________________\n",
      "bidirectional_12 (Bidirectio (None, 256)               140288    \n",
      "_________________________________________________________________\n",
      "dense_25 (Dense)             (None, 2)                 514       \n",
      "=================================================================\n",
      "Total params: 140,962\n",
      "Trainable params: 140,962\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "epochs = 100\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Embedding(20, 8, input_length=9))\n",
    "#model.add(Conv2D(1, kernel_size = (1,20), activation='relu', input_shape=(9,20,1)))\n",
    "#model.add(Flatten(data_format=None))\n",
    "#model.add(Dense(128, activation='relu'))\n",
    "#model.add(Dense(128, activation='sigmoid'))\n",
    "model.add(Bidirectional(LSTM(128, kernel_regularizer=regularizers.l2(0.1), recurrent_regularizer=regularizers.l2(0.1), \n",
    "                             bias_regularizer=regularizers.l2(0.1)), merge_mode='concat', weights=None))\n",
    "model.add(Dense(2, activation='softmax'))\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam',metrics=['accuracy'])\n",
    "\n",
    "model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 394 samples, validate on 99 samples\n",
      "Epoch 1/100\n",
      "394/394 [==============================] - 3s 9ms/step - loss: 49.7444 - acc: 0.5203 - val_loss: 43.6561 - val_acc: 0.5253\n",
      "Epoch 2/100\n",
      "394/394 [==============================] - 0s 341us/step - loss: 40.0382 - acc: 0.5457 - val_loss: 36.0092 - val_acc: 0.5253\n",
      "Epoch 3/100\n",
      "394/394 [==============================] - 0s 333us/step - loss: 33.6621 - acc: 0.5457 - val_loss: 31.0546 - val_acc: 0.5253\n",
      "Epoch 4/100\n",
      "394/394 [==============================] - 0s 360us/step - loss: 29.5242 - acc: 0.5457 - val_loss: 27.8083 - val_acc: 0.5253\n",
      "Epoch 5/100\n",
      "394/394 [==============================] - 0s 341us/step - loss: 26.7765 - acc: 0.5457 - val_loss: 25.6035 - val_acc: 0.5253\n",
      "Epoch 6/100\n",
      "394/394 [==============================] - 0s 326us/step - loss: 24.8757 - acc: 0.5457 - val_loss: 24.0335 - val_acc: 0.5253\n",
      "Epoch 7/100\n",
      "394/394 [==============================] - 0s 327us/step - loss: 23.4912 - acc: 0.5457 - val_loss: 22.8537 - val_acc: 0.5253\n",
      "Epoch 8/100\n",
      "394/394 [==============================] - 0s 332us/step - loss: 22.4254 - acc: 0.5457 - val_loss: 21.9140 - val_acc: 0.5253\n",
      "Epoch 9/100\n",
      "394/394 [==============================] - 0s 320us/step - loss: 21.5569 - acc: 0.5457 - val_loss: 21.1249 - val_acc: 0.5253\n",
      "Epoch 10/100\n",
      "394/394 [==============================] - 0s 369us/step - loss: 20.8146 - acc: 0.5457 - val_loss: 20.4333 - val_acc: 0.5253\n",
      "Epoch 11/100\n",
      "394/394 [==============================] - 0s 359us/step - loss: 20.1545 - acc: 0.5457 - val_loss: 19.8072 - val_acc: 0.5253\n",
      "Epoch 12/100\n",
      "394/394 [==============================] - 0s 376us/step - loss: 19.5467 - acc: 0.5457 - val_loss: 19.2222 - val_acc: 0.5253\n",
      "Epoch 13/100\n",
      "394/394 [==============================] - 0s 409us/step - loss: 18.9776 - acc: 0.5457 - val_loss: 18.6687 - val_acc: 0.5253\n",
      "Epoch 14/100\n",
      "394/394 [==============================] - 0s 400us/step - loss: 18.4338 - acc: 0.5457 - val_loss: 18.1374 - val_acc: 0.5253\n",
      "Epoch 15/100\n",
      "394/394 [==============================] - 0s 363us/step - loss: 17.9106 - acc: 0.5533 - val_loss: 17.6230 - val_acc: 0.6162\n",
      "Epoch 16/100\n",
      "394/394 [==============================] - 0s 316us/step - loss: 17.4039 - acc: 0.6777 - val_loss: 17.1232 - val_acc: 0.6768\n",
      "Epoch 17/100\n",
      "394/394 [==============================] - 0s 349us/step - loss: 16.9090 - acc: 0.6193 - val_loss: 16.6348 - val_acc: 0.6566\n",
      "Epoch 18/100\n",
      "394/394 [==============================] - 0s 359us/step - loss: 16.4237 - acc: 0.6193 - val_loss: 16.1547 - val_acc: 0.7071\n",
      "Epoch 19/100\n",
      "394/394 [==============================] - 0s 406us/step - loss: 15.9503 - acc: 0.7183 - val_loss: 15.6871 - val_acc: 0.7172\n",
      "Epoch 20/100\n",
      "394/394 [==============================] - 0s 405us/step - loss: 15.4852 - acc: 0.7234 - val_loss: 15.2298 - val_acc: 0.7374\n",
      "Epoch 21/100\n",
      "394/394 [==============================] - 0s 356us/step - loss: 15.0275 - acc: 0.7411 - val_loss: 14.7818 - val_acc: 0.7172\n",
      "Epoch 22/100\n",
      "394/394 [==============================] - 0s 344us/step - loss: 14.5835 - acc: 0.7462 - val_loss: 14.3460 - val_acc: 0.7475\n",
      "Epoch 23/100\n",
      "394/394 [==============================] - 0s 326us/step - loss: 14.1553 - acc: 0.7335 - val_loss: 13.9206 - val_acc: 0.7172\n",
      "Epoch 24/100\n",
      "394/394 [==============================] - 0s 379us/step - loss: 13.7326 - acc: 0.7513 - val_loss: 13.5090 - val_acc: 0.7273\n",
      "Epoch 25/100\n",
      "394/394 [==============================] - 0s 419us/step - loss: 13.3235 - acc: 0.7563 - val_loss: 13.1011 - val_acc: 0.7374\n",
      "Epoch 26/100\n",
      "394/394 [==============================] - 0s 412us/step - loss: 12.9281 - acc: 0.7716 - val_loss: 12.7063 - val_acc: 0.7374\n",
      "Epoch 27/100\n",
      "394/394 [==============================] - 0s 383us/step - loss: 12.5384 - acc: 0.7614 - val_loss: 12.3312 - val_acc: 0.7374\n",
      "Epoch 28/100\n",
      "394/394 [==============================] - 0s 317us/step - loss: 12.1598 - acc: 0.7538 - val_loss: 11.9575 - val_acc: 0.7475\n",
      "Epoch 29/100\n",
      "394/394 [==============================] - 0s 365us/step - loss: 11.7917 - acc: 0.7614 - val_loss: 11.5891 - val_acc: 0.7475\n",
      "Epoch 30/100\n",
      "394/394 [==============================] - 0s 327us/step - loss: 11.4318 - acc: 0.7487 - val_loss: 11.2345 - val_acc: 0.7475\n",
      "Epoch 31/100\n",
      "394/394 [==============================] - 0s 344us/step - loss: 11.0844 - acc: 0.7513 - val_loss: 10.8906 - val_acc: 0.7475\n",
      "Epoch 32/100\n",
      "394/394 [==============================] - 0s 372us/step - loss: 10.7400 - acc: 0.7513 - val_loss: 10.5562 - val_acc: 0.7374\n",
      "Epoch 33/100\n",
      "394/394 [==============================] - 0s 413us/step - loss: 10.4097 - acc: 0.7589 - val_loss: 10.2272 - val_acc: 0.7475\n",
      "Epoch 34/100\n",
      "394/394 [==============================] - 0s 410us/step - loss: 10.0834 - acc: 0.7690 - val_loss: 9.9206 - val_acc: 0.7374\n",
      "Epoch 35/100\n",
      "394/394 [==============================] - 0s 371us/step - loss: 9.7685 - acc: 0.7614 - val_loss: 9.5972 - val_acc: 0.7475\n",
      "Epoch 36/100\n",
      "394/394 [==============================] - 0s 346us/step - loss: 9.4617 - acc: 0.7563 - val_loss: 9.2961 - val_acc: 0.7475\n",
      "Epoch 37/100\n",
      "394/394 [==============================] - 0s 315us/step - loss: 9.1716 - acc: 0.7690 - val_loss: 9.0015 - val_acc: 0.7778\n",
      "Epoch 38/100\n",
      "394/394 [==============================] - 0s 307us/step - loss: 8.8841 - acc: 0.7437 - val_loss: 8.7211 - val_acc: 0.7374\n",
      "Epoch 39/100\n",
      "394/394 [==============================] - 0s 314us/step - loss: 8.5883 - acc: 0.7665 - val_loss: 8.4372 - val_acc: 0.7374\n",
      "Epoch 40/100\n",
      "394/394 [==============================] - 0s 310us/step - loss: 8.3229 - acc: 0.7538 - val_loss: 8.1779 - val_acc: 0.7475\n",
      "Epoch 41/100\n",
      "394/394 [==============================] - 0s 321us/step - loss: 8.0463 - acc: 0.7665 - val_loss: 7.8886 - val_acc: 0.7677\n",
      "Epoch 42/100\n",
      "394/394 [==============================] - 0s 394us/step - loss: 7.7834 - acc: 0.7563 - val_loss: 7.6482 - val_acc: 0.7475\n",
      "Epoch 43/100\n",
      "394/394 [==============================] - 0s 403us/step - loss: 7.5313 - acc: 0.7640 - val_loss: 7.3922 - val_acc: 0.7576\n",
      "Epoch 44/100\n",
      "394/394 [==============================] - 0s 398us/step - loss: 7.2835 - acc: 0.7614 - val_loss: 7.1574 - val_acc: 0.7374\n",
      "Epoch 45/100\n",
      "394/394 [==============================] - 0s 362us/step - loss: 7.0476 - acc: 0.7538 - val_loss: 6.9158 - val_acc: 0.7475\n",
      "Epoch 46/100\n",
      "394/394 [==============================] - 0s 399us/step - loss: 6.8116 - acc: 0.7665 - val_loss: 6.6827 - val_acc: 0.7576\n",
      "Epoch 47/100\n",
      "394/394 [==============================] - 0s 394us/step - loss: 6.5821 - acc: 0.7614 - val_loss: 6.4714 - val_acc: 0.7475\n",
      "Epoch 48/100\n",
      "394/394 [==============================] - 0s 333us/step - loss: 6.3632 - acc: 0.7640 - val_loss: 6.2474 - val_acc: 0.7374\n",
      "Epoch 49/100\n",
      "394/394 [==============================] - 0s 358us/step - loss: 6.1477 - acc: 0.7538 - val_loss: 6.0364 - val_acc: 0.7374\n",
      "Epoch 50/100\n",
      "394/394 [==============================] - 0s 362us/step - loss: 5.9457 - acc: 0.7614 - val_loss: 5.8380 - val_acc: 0.7374\n",
      "Epoch 51/100\n",
      "394/394 [==============================] - 0s 376us/step - loss: 5.7451 - acc: 0.7513 - val_loss: 5.6332 - val_acc: 0.7374\n",
      "Epoch 52/100\n",
      "394/394 [==============================] - 0s 353us/step - loss: 5.5418 - acc: 0.7640 - val_loss: 5.4467 - val_acc: 0.7374\n",
      "Epoch 53/100\n",
      "394/394 [==============================] - 0s 357us/step - loss: 5.3637 - acc: 0.7513 - val_loss: 5.2645 - val_acc: 0.7374\n",
      "Epoch 54/100\n",
      "394/394 [==============================] - 0s 360us/step - loss: 5.1814 - acc: 0.7589 - val_loss: 5.0919 - val_acc: 0.7374\n",
      "Epoch 55/100\n",
      "394/394 [==============================] - 0s 304us/step - loss: 4.9913 - acc: 0.7589 - val_loss: 4.8887 - val_acc: 0.7677\n",
      "Epoch 56/100\n",
      "394/394 [==============================] - 0s 391us/step - loss: 4.8232 - acc: 0.7640 - val_loss: 4.7356 - val_acc: 0.7374\n",
      "Epoch 57/100\n",
      "394/394 [==============================] - 0s 378us/step - loss: 4.6499 - acc: 0.7741 - val_loss: 4.5684 - val_acc: 0.7475\n",
      "Epoch 58/100\n",
      "394/394 [==============================] - 0s 369us/step - loss: 4.4876 - acc: 0.7640 - val_loss: 4.4019 - val_acc: 0.7576\n",
      "Epoch 59/100\n",
      "394/394 [==============================] - 0s 319us/step - loss: 4.3314 - acc: 0.7690 - val_loss: 4.2573 - val_acc: 0.7374\n",
      "Epoch 60/100\n",
      "394/394 [==============================] - 0s 308us/step - loss: 4.1803 - acc: 0.7690 - val_loss: 4.1080 - val_acc: 0.7475\n",
      "Epoch 61/100\n",
      "394/394 [==============================] - 0s 309us/step - loss: 4.0386 - acc: 0.7716 - val_loss: 3.9637 - val_acc: 0.7475\n",
      "Epoch 62/100\n",
      "394/394 [==============================] - 0s 306us/step - loss: 3.8929 - acc: 0.7589 - val_loss: 3.8263 - val_acc: 0.7273\n",
      "Epoch 63/100\n",
      "394/394 [==============================] - 0s 310us/step - loss: 3.7574 - acc: 0.7741 - val_loss: 3.6814 - val_acc: 0.7475\n",
      "Epoch 64/100\n",
      "394/394 [==============================] - 0s 351us/step - loss: 3.6204 - acc: 0.7614 - val_loss: 3.5681 - val_acc: 0.7374\n",
      "Epoch 65/100\n",
      "394/394 [==============================] - 0s 350us/step - loss: 3.4891 - acc: 0.7716 - val_loss: 3.4234 - val_acc: 0.7475\n",
      "Epoch 66/100\n",
      "394/394 [==============================] - 0s 336us/step - loss: 3.3694 - acc: 0.7614 - val_loss: 3.3146 - val_acc: 0.7475\n",
      "Epoch 67/100\n",
      "394/394 [==============================] - 0s 303us/step - loss: 3.2423 - acc: 0.7716 - val_loss: 3.1851 - val_acc: 0.7475\n",
      "Epoch 68/100\n",
      "394/394 [==============================] - 0s 299us/step - loss: 3.1269 - acc: 0.7741 - val_loss: 3.0826 - val_acc: 0.7475\n",
      "Epoch 69/100\n",
      "394/394 [==============================] - 0s 305us/step - loss: 3.0154 - acc: 0.7792 - val_loss: 2.9627 - val_acc: 0.7475\n",
      "Epoch 70/100\n",
      "394/394 [==============================] - 0s 354us/step - loss: 2.9174 - acc: 0.7589 - val_loss: 2.8609 - val_acc: 0.7475\n",
      "Epoch 71/100\n",
      "394/394 [==============================] - 0s 339us/step - loss: 2.8043 - acc: 0.7690 - val_loss: 2.7618 - val_acc: 0.7475\n",
      "Epoch 72/100\n",
      "394/394 [==============================] - 0s 337us/step - loss: 2.6994 - acc: 0.7766 - val_loss: 2.6644 - val_acc: 0.7475\n",
      "Epoch 73/100\n",
      "394/394 [==============================] - 0s 327us/step - loss: 2.6057 - acc: 0.7741 - val_loss: 2.5671 - val_acc: 0.7475\n",
      "Epoch 74/100\n",
      "394/394 [==============================] - 0s 321us/step - loss: 2.5152 - acc: 0.7589 - val_loss: 2.4802 - val_acc: 0.7576\n",
      "Epoch 75/100\n",
      "394/394 [==============================] - 0s 336us/step - loss: 2.4297 - acc: 0.7640 - val_loss: 2.3930 - val_acc: 0.7475\n",
      "Epoch 76/100\n",
      "394/394 [==============================] - 0s 379us/step - loss: 2.3360 - acc: 0.7766 - val_loss: 2.3010 - val_acc: 0.7576\n",
      "Epoch 77/100\n",
      "394/394 [==============================] - 0s 319us/step - loss: 2.2473 - acc: 0.7741 - val_loss: 2.2315 - val_acc: 0.7475\n",
      "Epoch 78/100\n",
      "394/394 [==============================] - 0s 322us/step - loss: 2.1695 - acc: 0.7665 - val_loss: 2.1468 - val_acc: 0.7374\n",
      "Epoch 79/100\n",
      "394/394 [==============================] - 0s 339us/step - loss: 2.0934 - acc: 0.7716 - val_loss: 2.0648 - val_acc: 0.7677\n",
      "Epoch 80/100\n",
      "394/394 [==============================] - 0s 350us/step - loss: 2.0191 - acc: 0.7817 - val_loss: 1.9940 - val_acc: 0.7778\n",
      "Epoch 81/100\n",
      "394/394 [==============================] - 0s 348us/step - loss: 1.9531 - acc: 0.7640 - val_loss: 1.9294 - val_acc: 0.7475\n",
      "Epoch 82/100\n",
      "394/394 [==============================] - 0s 372us/step - loss: 1.8774 - acc: 0.7741 - val_loss: 1.8645 - val_acc: 0.7677\n",
      "Epoch 83/100\n",
      "394/394 [==============================] - 0s 370us/step - loss: 1.8167 - acc: 0.7589 - val_loss: 1.7964 - val_acc: 0.7576\n",
      "Epoch 84/100\n",
      "394/394 [==============================] - 0s 346us/step - loss: 1.7509 - acc: 0.7741 - val_loss: 1.7345 - val_acc: 0.7576\n",
      "Epoch 85/100\n",
      "394/394 [==============================] - 0s 333us/step - loss: 1.6862 - acc: 0.7766 - val_loss: 1.6729 - val_acc: 0.7778\n",
      "Epoch 86/100\n",
      "394/394 [==============================] - 0s 367us/step - loss: 1.6258 - acc: 0.7868 - val_loss: 1.6147 - val_acc: 0.7677\n",
      "Epoch 87/100\n",
      "394/394 [==============================] - 0s 368us/step - loss: 1.5698 - acc: 0.7766 - val_loss: 1.5664 - val_acc: 0.7475\n",
      "Epoch 88/100\n",
      "394/394 [==============================] - 0s 345us/step - loss: 1.5183 - acc: 0.7792 - val_loss: 1.5166 - val_acc: 0.7677\n",
      "Epoch 89/100\n",
      "394/394 [==============================] - 0s 328us/step - loss: 1.4715 - acc: 0.7792 - val_loss: 1.4628 - val_acc: 0.7879\n",
      "Epoch 90/100\n",
      "394/394 [==============================] - 0s 352us/step - loss: 1.4182 - acc: 0.7919 - val_loss: 1.4113 - val_acc: 0.7677\n",
      "Epoch 91/100\n",
      "394/394 [==============================] - 0s 386us/step - loss: 1.3821 - acc: 0.7741 - val_loss: 1.3837 - val_acc: 0.7374\n",
      "Epoch 92/100\n",
      "394/394 [==============================] - 0s 378us/step - loss: 1.3461 - acc: 0.7513 - val_loss: 1.3187 - val_acc: 0.7980\n",
      "Epoch 93/100\n",
      "394/394 [==============================] - 0s 336us/step - loss: 1.2892 - acc: 0.7640 - val_loss: 1.2957 - val_acc: 0.7576\n",
      "Epoch 94/100\n",
      "394/394 [==============================] - 0s 326us/step - loss: 1.2507 - acc: 0.7716 - val_loss: 1.2418 - val_acc: 0.7677\n",
      "Epoch 95/100\n",
      "394/394 [==============================] - 0s 349us/step - loss: 1.2019 - acc: 0.7665 - val_loss: 1.2041 - val_acc: 0.7778\n",
      "Epoch 96/100\n",
      "394/394 [==============================] - 0s 334us/step - loss: 1.1615 - acc: 0.7792 - val_loss: 1.1758 - val_acc: 0.7879\n",
      "Epoch 97/100\n",
      "394/394 [==============================] - 0s 373us/step - loss: 1.1380 - acc: 0.7563 - val_loss: 1.1276 - val_acc: 0.7879\n",
      "Epoch 98/100\n",
      "394/394 [==============================] - 0s 382us/step - loss: 1.1185 - acc: 0.7589 - val_loss: 1.1129 - val_acc: 0.7475\n",
      "Epoch 99/100\n",
      "394/394 [==============================] - 0s 373us/step - loss: 1.0691 - acc: 0.7792 - val_loss: 1.0596 - val_acc: 0.8182\n",
      "Epoch 100/100\n",
      "394/394 [==============================] - 0s 362us/step - loss: 1.0298 - acc: 0.7843 - val_loss: 1.0639 - val_acc: 0.7475\n"
     ]
    }
   ],
   "source": [
    "model_train = model.fit(train_X, train_label, epochs=epochs, batch_size=32, validation_data=(valid_X, valid_label))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the training accuracy\n",
    "accuracy = model_train.history['acc']\n",
    "val_accuracy = model_train.history['val_acc']\n",
    "loss = model_train.history['loss']\n",
    "val_loss = model_train.history['val_loss']\n",
    "epochs = range(len(accuracy))\n",
    "plt.plot(epochs, accuracy, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_accuracy, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}