2660 lines (2659 with data), 515.6 kB

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ajHeD1nTIB1j"
   },
   "source": [
    "# 1DCNN code for damage localization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Mjdojg-1ILA4"
   },
   "source": [
    "## Importing APIs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "IkeIr5YWbktM"
   },
   "outputs": [],
   "source": [
    "#imports\n",
    "import os\n",
    "import zipfile\n",
    "%matplotlib inline\n",
    "import random \n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Flatten, Activation, Conv1D, MaxPooling1D, Dropout, Lambda, BatchNormalization\n",
    "from tensorflow.keras.optimizers import SGD, Adam, RMSprop\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import preprocessing\n",
    "from keras.utils import to_categorical,plot_model\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "from scipy import stats\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.compat.v1 import ConfigProto\n",
    "from tensorflow.compat.v1 import InteractiveSession\n",
    "config = ConfigProto()\n",
    "config.gpu_options.allow_growth = True\n",
    "session = InteractiveSession(config=config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "7ildUOZpINr6"
   },
   "source": [
    "## Callbacks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "H-lBgT16UzvV"
   },
   "outputs": [],
   "source": [
    "class myCallback(tf.keras.callbacks.Callback):\n",
    "  def on_epoch_end(self, epoch, logs={}):\n",
    "    if (logs.get('val_acc')>0.99) and (logs.get('acc')>0.99) and (logs.get('val_loss')<0.07) and (logs.get('loss')<0.07):\n",
    "      print(\"\\nReached perfect accuracy so cancelling training!\")\n",
    "      self.model.stop_training = True\n",
    "\n",
    "epoch_schedule = myCallback()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "50uuiQSSyqyj"
   },
   "outputs": [],
   "source": [
    "lr_schedule = tf.keras.callbacks.LearningRateScheduler(\n",
    "    lambda epoch: 1e-6 * 10**(epoch / 5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "gPvQTFpd5iZv"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(22176, 13108)\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv('E:/PhD-MSR/OGWdataset/Guided_wave_basic_measurement_data/dataset/AllDatasets/combinedata/SHM_diff_filter.txt',header=None)\n",
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "t9btG2kdIU6H"
   },
   "source": [
    "## Plotting dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 12637,
     "status": "ok",
     "timestamp": 1571626983217,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "WK6ayI--9vlk",
    "outputId": "f2f7e207-04b5-4b05-e56c-6587ffec9b9f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(13108,)\n"
     ]
    }
   ],
   "source": [
    "seqlen = df.shape[1]\n",
    "dT = 1e-7\n",
    "time = np.arange(0,(seqlen)*dT,dT)\n",
    "print(time.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "sigwindow = 1e-3\n",
    "idx1 = 1000\n",
    "idx2 = 6100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 316
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 12336,
     "status": "ok",
     "timestamp": 1571626983575,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "GzmE3vxpFo4Z",
    "outputId": "c4d9037a-65c0-4119-b374-aa1dd3dade4a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7141\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Norm Amplitude')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,4))\n",
    "nn = random.randint(1,df.shape[0])\n",
    "plt.plot(time[idx1:idx2], df.iloc[nn,idx1: idx2])\n",
    "plt.title('Training example',fontsize=25)\n",
    "plt.xticks(fontsize=25)\n",
    "plt.yticks(fontsize=25)\n",
    "print(nn)\n",
    "plt.xlabel('Time is seconds',fontsize=25)\n",
    "plt.ylabel('Norm Amplitude',fontsize=25)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "CXy8CBXkIY0p"
   },
   "source": [
    "## Deciding Features and Labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# to save RAM\n",
    "#del freqtransch,transch,damtransch,A,B,X,y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3024,)\n",
      "[    6     7     8 ... 21565 21566 21567]\n"
     ]
    }
   ],
   "source": [
    "# Using only transmission channels\n",
    "sch1 = np.arange(6,11+1,1)                   # sub channels with 1 as actuator and other as sensors\n",
    "sch2 = np.arange(16,21+1,1)                  # sub channels with 2 as actuator and other as sensors\n",
    "sch3 = np.arange(25,30+1,1)\n",
    "sch4 = np.arange(33,38+1,1)\n",
    "sch5 = np.arange(40,45+1,1)\n",
    "sch6 = np.arange(46,51+1,1)\n",
    "transch = np.concatenate([sch1,sch2,sch3,sch4,sch5,sch6], axis=0)\n",
    "freqtransch = np.concatenate([transch,transch+66,transch+66*2], axis=0)\n",
    "\n",
    "A = []\n",
    "for i in range(1,28,1):                       # we are incorporating all damages, i = 1 here means D2\n",
    "    damtransch = 792*i + freqtransch      \n",
    "    A.append(damtransch)\n",
    "\n",
    "A = np.concatenate(A)                         # concatenate all appended vectors\n",
    "B = np.concatenate([freqtransch,A],axis=0)    # concatenate D1 with D2 to D28\n",
    "print(B.shape)\n",
    "print(B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Mn_gsCMfVY6N"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3024, 5100)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Input/Features and labels extraction\n",
    "X = df.iloc[B, idx1:idx2]\n",
    "X = np.array(X)\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 316
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 12336,
     "status": "ok",
     "timestamp": 1571626983575,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "GzmE3vxpFo4Z",
    "outputId": "c4d9037a-65c0-4119-b374-aa1dd3dade4a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1816\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Norm Amplitude')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,4))\n",
    "nn = random.randint(1,X.shape[0])\n",
    "plt.plot(time[idx1:idx2], X[nn,0:X.shape[1]])\n",
    "plt.title('Training example',fontsize=25)\n",
    "plt.xticks(fontsize=25)\n",
    "plt.yticks(fontsize=25)\n",
    "print(nn)\n",
    "plt.xlabel('Time is seconds',fontsize=25)\n",
    "plt.ylabel('Norm Amplitude',fontsize=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3024, 7)\n"
     ]
    }
   ],
   "source": [
    "# Input/Features and labels extraction\n",
    "# Baseline = 0\n",
    "# Damage = 1\n",
    "mclass = freqtransch.shape[0]*4\n",
    "y1 = np.zeros((mclass,1), dtype=int)\n",
    "y2 = np.ones((mclass,1), dtype=int)\n",
    "y3 = y2*2\n",
    "y4 = y2*3\n",
    "y5 = y2*4\n",
    "y6 = y2*5\n",
    "y7 = y2*6\n",
    "\n",
    "y = np.concatenate([y1,y2,y3,y4,y5,y6,y7], axis=0)\n",
    "y = np.array(y)\n",
    "\n",
    "y = to_categorical(y)\n",
    "print(y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#del Xn1,n1,r,maxX"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 5100)\n",
      "()\n",
      "(1, 5100)\n",
      "(3024, 5100)\n",
      "4.905745337030785e-06\n",
      "2.447573484037756e-08\n",
      "SNR :  23.01969260132006\n"
     ]
    }
   ],
   "source": [
    "#---Random gaussian noise parameter\n",
    "beta1 = 0.0025\n",
    "beta2 = 0.003\n",
    "\n",
    "mu = 0\n",
    "sigma = 1\n",
    "\n",
    "r = sigma*np.random.randn(X.shape[1],1) + mu   #random parameter with gaussian distribution\n",
    "r = np.transpose(r)\n",
    "print(r.shape)\n",
    "\n",
    "#---Noisy signal\n",
    "maxX = np.max(X)\n",
    "maxX = np.array(maxX)\n",
    "print(maxX.shape)\n",
    "\n",
    "n1 = beta1*r*maxX\n",
    "n2 = beta2*r*maxX\n",
    "print(n2.shape)\n",
    "\n",
    "Xn1 = X + n1\n",
    "Xn2 = X + n2\n",
    "print(Xn2.shape)\n",
    "\n",
    "#---Signal to noise ratio\n",
    "import math\n",
    "\n",
    "rms_Xn = np.sqrt(np.mean(Xn2**2))\n",
    "Power_Xn = rms_Xn**2\n",
    "print(Power_Xn)\n",
    "\n",
    "rms_n = np.sqrt(np.mean(n2**2))\n",
    "Power_n = rms_n**2\n",
    "print(Power_n)\n",
    "\n",
    "SNR_dB = 10*math.log10(Power_Xn/Power_n)\n",
    "print(\"SNR : \",SNR_dB)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(9072, 5100)\n",
      "(9072, 7)\n"
     ]
    }
   ],
   "source": [
    "Xn = np.concatenate([X,Xn1,Xn2], axis=0)\n",
    "yn = np.concatenate([y,y,y], axis=0)\n",
    "print(Xn.shape)\n",
    "print(yn.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8253\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Dam')"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#---Plot non-noisy signal\n",
    "D = random.randint(1, Xn.shape[0])\n",
    "print(D)\n",
    "\n",
    "plt.figure(figsize=(10,4))\n",
    "plt.plot(time[idx1:idx2], Xn[D, 0 : Xn.shape[1]])\n",
    "plt.title('Dam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 9748,
     "status": "ok",
     "timestamp": 1571626984031,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "y6KgAnxXVbNQ",
    "outputId": "f3c40a74-246b-4f95-d607-15acf4650c55"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(9072, 5100)\n",
      "(9072, 7)\n"
     ]
    }
   ],
   "source": [
    "# Shapes of input and output\n",
    "print(Xn.shape)\n",
    "print(yn.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# https://towardsdatascience.com/a-bunch-of-tips-and-tricks-for-training-deep-neural-networks-3ca24c31ddc8"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "_Ga43nEOId8F"
   },
   "source": [
    "## Splitting testing and test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#del model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GXBYgxIbVdaH"
   },
   "outputs": [],
   "source": [
    "# split into train and test\n",
    "# Run this line of code for every run of the code if some parameters are changed\n",
    "X1, X_test, y1, y_test = train_test_split(Xn,yn, test_size = 0.0005, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_valid, y_train, y_valid = train_test_split(X1,y1, test_size = 0.1, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 87
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 7372,
     "status": "ok",
     "timestamp": 1571626984037,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "qtkixnevVfr4",
    "outputId": "3c7ae0b4-d2fb-4bd8-b7e1-7c6ed8a0dedb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8160, 5100)\n",
      "(907, 5100)\n",
      "(5, 5100)\n",
      "(8160, 7)\n",
      "(907, 7)\n",
      "(5, 7)\n"
     ]
    }
   ],
   "source": [
    "# Shapes of training and test tests\n",
    "X_train = np.array(X_train)\n",
    "X_valid = np.array(X_valid)\n",
    "X_test = np.array(X_test)\n",
    "y_train = np.array(y_train)\n",
    "y_train = np.array(y_train)\n",
    "y_train = np.array(y_train)\n",
    "print(X_train.shape)\n",
    "print(X_valid.shape)\n",
    "print(X_test.shape)\n",
    "print(y_train.shape)\n",
    "print(y_valid.shape)\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "N3mc165NIib0"
   },
   "source": [
    "## Reshape the arrays"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 212
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 6545,
     "status": "ok",
     "timestamp": 1571626984335,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "S0GkyLE2ViB4",
    "outputId": "bd475f73-8cdc-444e-e576-87df1bd9d07c"
   },
   "outputs": [],
   "source": [
    "# reshape the arrays\n",
    "X_train = np.reshape(X_train, (X_train.shape[0],X_train.shape[1],1))\n",
    "X_valid = np.reshape(X_valid, (X_valid.shape[0],X_valid.shape[1],1))\n",
    "X_test  = np.reshape(X_test,  (X_test.shape[0],X_test.shape[1],1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8160, 5100, 1)\n",
      "(907, 5100, 1)\n",
      "(8160, 7)\n",
      "(907, 7)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape)\n",
    "print(X_valid.shape)\n",
    "print(y_train.shape)\n",
    "print(y_valid.shape)\n",
    "# print(y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "XW7VmHZfInYs"
   },
   "source": [
    "## Model Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 90
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 825,
     "status": "ok",
     "timestamp": 1571626986180,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "azmX3H0DVkkJ",
    "outputId": "73d34f6a-75d7-423e-a653-87a636ea6ef8"
   },
   "outputs": [],
   "source": [
    "# define model architecture : 1DCNN-classification\n",
    "from keras.layers import Input\n",
    "from keras.models import Model\n",
    "from keras.layers import concatenate\n",
    "from keras.models import Sequential\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.layers.convolutional import Conv1D\n",
    "from keras.layers.convolutional import MaxPooling1D\n",
    "from keras.layers.core import Activation\n",
    "from keras.layers.core import Dropout\n",
    "from keras.layers.core import Dense\n",
    "from keras.layers import Flatten\n",
    "from keras.layers import Input\n",
    "from keras.models import Model\n",
    "\n",
    "Inp = Input(shape=(X.shape[1],1))\n",
    "\n",
    "# Here i am trying multiscale CNN\n",
    "# scale 1 = 3, scale 2 = 7\n",
    "x1 = Sequential()\n",
    "\n",
    "x1 = Conv1D(filters=16, kernel_size=3, input_shape=(X.shape[1],1))(Inp)\n",
    "x1 = MaxPooling1D(pool_size=2)(x1)\n",
    "\n",
    "x1 = Conv1D(filters=32, kernel_size=3, activation='relu')(x1)\n",
    "x1 = MaxPooling1D(pool_size=2)(x1)\n",
    "\n",
    "x1 = Conv1D(filters=64, kernel_size=3, activation='relu')(x1)\n",
    "x1 = MaxPooling1D(pool_size=2)(x1)\n",
    "\n",
    "x1 = Conv1D(filters=128, kernel_size=3, activation='relu')(x1)\n",
    "x1 = MaxPooling1D(pool_size=2)(x1)\n",
    "\n",
    "x1 = Conv1D(filters=256, kernel_size=3, activation='relu')(x1)\n",
    "x1 = MaxPooling1D(pool_size=2)(x1)\n",
    "\n",
    "x1 = Flatten()(x1)\n",
    "\n",
    "cnn1 = Model(Inp,x1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.layers import concatenate\n",
    "#x = concatenate([cnn1.output,cnn2.output])\n",
    "x = cnn1.output\n",
    "#---------------------------------\n",
    "x = Dense(128, activation='relu')(x)\n",
    "x = Dropout(0.25)(x)\n",
    "x = Dense(7, activation='softmax')(x)\n",
    "#---------------------------------\n",
    "model = Model(inputs = Inp, outputs=x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 478
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 559,
     "status": "ok",
     "timestamp": 1571626986708,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "JMZwl4oJVm_s",
    "outputId": "049265c3-1b4f-4ff1-a307-c5862f1ffd09"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_6\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_3 (InputLayer)         (None, 5100, 1)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_11 (Conv1D)           (None, 5098, 16)          64        \n",
      "_________________________________________________________________\n",
      "max_pooling1d_11 (MaxPooling (None, 2549, 16)          0         \n",
      "_________________________________________________________________\n",
      "conv1d_12 (Conv1D)           (None, 2547, 32)          1568      \n",
      "_________________________________________________________________\n",
      "max_pooling1d_12 (MaxPooling (None, 1273, 32)          0         \n",
      "_________________________________________________________________\n",
      "conv1d_13 (Conv1D)           (None, 1271, 64)          6208      \n",
      "_________________________________________________________________\n",
      "max_pooling1d_13 (MaxPooling (None, 635, 64)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_14 (Conv1D)           (None, 633, 128)          24704     \n",
      "_________________________________________________________________\n",
      "max_pooling1d_14 (MaxPooling (None, 316, 128)          0         \n",
      "_________________________________________________________________\n",
      "conv1d_15 (Conv1D)           (None, 314, 256)          98560     \n",
      "_________________________________________________________________\n",
      "max_pooling1d_15 (MaxPooling (None, 157, 256)          0         \n",
      "_________________________________________________________________\n",
      "flatten_3 (Flatten)          (None, 40192)             0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 128)               5144704   \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 7)                 903       \n",
      "=================================================================\n",
      "Total params: 5,276,711\n",
      "Trainable params: 5,276,711\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# summary of the model\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import keras\n",
    "keras.utils.plot_model(model, \"my_first_model.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Functions for recall, precision and f1 score\n",
    "from keras import backend as K\n",
    "\n",
    "def recall_m(y_true, y_pred):\n",
    "    true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
    "    possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))\n",
    "    recall = true_positives / (possible_positives + K.epsilon())\n",
    "    return recall\n",
    "\n",
    "def precision_m(y_true, y_pred):\n",
    "    true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
    "    predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))\n",
    "    precision = true_positives / (predicted_positives + K.epsilon())\n",
    "    return precision\n",
    "\n",
    "def f1_m(y_true, y_pred):\n",
    "    precision = precision_m(y_true, y_pred)\n",
    "    recall = recall_m(y_true, y_pred)\n",
    "    return 2*((precision*recall)/(precision+recall+K.epsilon()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "8tiR31NyIqjJ"
   },
   "source": [
    "## Choose Hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 90
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 821,
     "status": "ok",
     "timestamp": 1571626989175,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "3TxHGquwVpZO",
    "outputId": "f8f6c7a4-e85c-4c76-c5da-71a7589ab847"
   },
   "outputs": [],
   "source": [
    "# compile model\n",
    "model.compile(loss='categorical_crossentropy', optimizer=Adam(lr=1e-5), metrics=['acc', precision_m, recall_m])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SAqS8ppy-13X"
   },
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 887
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 73050,
     "status": "ok",
     "timestamp": 1571627063426,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "v6kF1cwfVu0n",
    "outputId": "ada6c377-ce02-4132-a01a-17832050313d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 8160 samples, validate on 907 samples\n",
      "Epoch 1/1000\n",
      " - 3s - loss: 1.9459 - acc: 0.1441 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9458 - val_acc: 0.1367 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 2/1000\n",
      " - 2s - loss: 1.9455 - acc: 0.1587 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9454 - val_acc: 0.1378 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 3/1000\n",
      " - 2s - loss: 1.9446 - acc: 0.1561 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9439 - val_acc: 0.1676 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 4/1000\n",
      " - 2s - loss: 1.9422 - acc: 0.1697 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9408 - val_acc: 0.1830 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 5/1000\n",
      " - 2s - loss: 1.9374 - acc: 0.1799 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9356 - val_acc: 0.1918 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 6/1000\n",
      " - 2s - loss: 1.9311 - acc: 0.1909 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9287 - val_acc: 0.2117 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 7/1000\n",
      " - 2s - loss: 1.9231 - acc: 0.2093 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9210 - val_acc: 0.2051 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 8/1000\n",
      " - 2s - loss: 1.9151 - acc: 0.2107 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9142 - val_acc: 0.2106 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 9/1000\n",
      " - 2s - loss: 1.9084 - acc: 0.2151 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9082 - val_acc: 0.2084 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 10/1000\n",
      " - 2s - loss: 1.9026 - acc: 0.2082 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.9037 - val_acc: 0.1963 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
      "Epoch 11/1000\n",
      " - 2s - loss: 1.8961 - acc: 0.2173 - precision_m: 0.0469 - recall_m: 7.3242e-04 - val_loss: 1.8967 - val_acc: 0.2183 - val_precision_m: 0.0667 - val_recall_m: 0.0010\n",
      "Epoch 12/1000\n",
      " - 2s - loss: 1.8889 - acc: 0.2205 - precision_m: 0.1016 - recall_m: 0.0016 - val_loss: 1.8912 - val_acc: 0.2183 - val_precision_m: 0.0667 - val_recall_m: 0.0010\n",
      "Epoch 13/1000\n",
      " - 2s - loss: 1.8840 - acc: 0.2246 - precision_m: 0.1250 - recall_m: 0.0020 - val_loss: 1.8843 - val_acc: 0.2249 - val_precision_m: 0.0667 - val_recall_m: 0.0010\n",
      "Epoch 14/1000\n",
      " - 2s - loss: 1.8767 - acc: 0.2301 - precision_m: 0.1484 - recall_m: 0.0027 - val_loss: 1.8784 - val_acc: 0.2227 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 15/1000\n",
      " - 2s - loss: 1.8702 - acc: 0.2364 - precision_m: 0.2266 - recall_m: 0.0035 - val_loss: 1.8726 - val_acc: 0.2393 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 16/1000\n",
      " - 2s - loss: 1.8629 - acc: 0.2387 - precision_m: 0.2383 - recall_m: 0.0042 - val_loss: 1.8653 - val_acc: 0.2359 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 17/1000\n",
      " - 2s - loss: 1.8575 - acc: 0.2371 - precision_m: 0.2578 - recall_m: 0.0045 - val_loss: 1.8605 - val_acc: 0.2348 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 18/1000\n",
      " - 2s - loss: 1.8493 - acc: 0.2445 - precision_m: 0.2656 - recall_m: 0.0050 - val_loss: 1.8544 - val_acc: 0.2426 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 19/1000\n",
      " - 2s - loss: 1.8448 - acc: 0.2442 - precision_m: 0.2943 - recall_m: 0.0060 - val_loss: 1.8475 - val_acc: 0.2415 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 20/1000\n",
      " - 2s - loss: 1.8364 - acc: 0.2523 - precision_m: 0.3086 - recall_m: 0.0067 - val_loss: 1.8416 - val_acc: 0.2337 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 21/1000\n",
      " - 2s - loss: 1.8284 - acc: 0.2561 - precision_m: 0.3633 - recall_m: 0.0072 - val_loss: 1.8373 - val_acc: 0.2415 - val_precision_m: 0.1333 - val_recall_m: 0.0021\n",
      "Epoch 22/1000\n",
      " - 2s - loss: 1.8196 - acc: 0.2585 - precision_m: 0.3685 - recall_m: 0.0083 - val_loss: 1.8298 - val_acc: 0.2370 - val_precision_m: 0.2000 - val_recall_m: 0.0031\n",
      "Epoch 23/1000\n",
      " - 2s - loss: 1.8124 - acc: 0.2616 - precision_m: 0.4570 - recall_m: 0.0089 - val_loss: 1.8229 - val_acc: 0.2426 - val_precision_m: 0.3333 - val_recall_m: 0.0052\n",
      "Epoch 24/1000\n",
      " - 2s - loss: 1.8047 - acc: 0.2640 - precision_m: 0.4635 - recall_m: 0.0107 - val_loss: 1.8158 - val_acc: 0.2481 - val_precision_m: 0.4000 - val_recall_m: 0.0063\n",
      "Epoch 25/1000\n",
      " - 2s - loss: 1.7956 - acc: 0.2681 - precision_m: 0.5102 - recall_m: 0.0123 - val_loss: 1.8107 - val_acc: 0.2503 - val_precision_m: 0.4000 - val_recall_m: 0.0063\n",
      "Epoch 26/1000\n",
      " - 2s - loss: 1.7890 - acc: 0.2765 - precision_m: 0.5788 - recall_m: 0.0128 - val_loss: 1.8058 - val_acc: 0.2602 - val_precision_m: 0.4333 - val_recall_m: 0.0083\n",
      "Epoch 27/1000\n",
      " - 2s - loss: 1.7805 - acc: 0.2761 - precision_m: 0.5017 - recall_m: 0.0135 - val_loss: 1.7961 - val_acc: 0.2503 - val_precision_m: 0.4333 - val_recall_m: 0.0083\n",
      "Epoch 28/1000\n",
      " - 2s - loss: 1.7680 - acc: 0.2848 - precision_m: 0.6367 - recall_m: 0.0170 - val_loss: 1.7912 - val_acc: 0.2525 - val_precision_m: 0.5000 - val_recall_m: 0.0094\n",
      "Epoch 29/1000\n",
      " - 2s - loss: 1.7646 - acc: 0.2859 - precision_m: 0.6677 - recall_m: 0.0178 - val_loss: 1.7857 - val_acc: 0.2558 - val_precision_m: 0.5000 - val_recall_m: 0.0144\n",
      "Epoch 30/1000\n",
      " - 2s - loss: 1.7514 - acc: 0.2900 - precision_m: 0.7221 - recall_m: 0.0203 - val_loss: 1.7782 - val_acc: 0.2657 - val_precision_m: 0.4667 - val_recall_m: 0.0144\n",
      "Epoch 31/1000\n",
      " - 2s - loss: 1.7471 - acc: 0.2934 - precision_m: 0.6401 - recall_m: 0.0217 - val_loss: 1.7718 - val_acc: 0.2646 - val_precision_m: 0.6667 - val_recall_m: 0.0186\n",
      "Epoch 32/1000\n",
      " - 2s - loss: 1.7396 - acc: 0.2934 - precision_m: 0.6695 - recall_m: 0.0232 - val_loss: 1.7763 - val_acc: 0.2679 - val_precision_m: 0.7444 - val_recall_m: 0.0238\n",
      "Epoch 33/1000\n",
      " - 2s - loss: 1.7319 - acc: 0.3067 - precision_m: 0.6227 - recall_m: 0.0242 - val_loss: 1.7602 - val_acc: 0.2646 - val_precision_m: 0.7444 - val_recall_m: 0.0248\n",
      "Epoch 34/1000\n",
      " - 2s - loss: 1.7252 - acc: 0.3081 - precision_m: 0.7303 - recall_m: 0.0260 - val_loss: 1.7548 - val_acc: 0.2745 - val_precision_m: 0.7444 - val_recall_m: 0.0259\n",
      "Epoch 35/1000\n",
      " - 2s - loss: 1.7220 - acc: 0.3082 - precision_m: 0.6828 - recall_m: 0.0273 - val_loss: 1.7479 - val_acc: 0.2723 - val_precision_m: 0.7444 - val_recall_m: 0.0248\n",
      "Epoch 36/1000\n",
      " - 2s - loss: 1.7078 - acc: 0.3167 - precision_m: 0.7281 - recall_m: 0.0314 - val_loss: 1.7399 - val_acc: 0.2668 - val_precision_m: 0.8111 - val_recall_m: 0.0269\n",
      "Epoch 37/1000\n",
      " - 2s - loss: 1.6991 - acc: 0.3190 - precision_m: 0.7681 - recall_m: 0.0327 - val_loss: 1.7365 - val_acc: 0.2778 - val_precision_m: 0.7500 - val_recall_m: 0.0248\n",
      "Epoch 38/1000\n",
      " - 2s - loss: 1.6909 - acc: 0.3241 - precision_m: 0.7063 - recall_m: 0.0354 - val_loss: 1.7321 - val_acc: 0.2856 - val_precision_m: 0.8556 - val_recall_m: 0.0331\n",
      "Epoch 39/1000\n",
      " - 2s - loss: 1.6879 - acc: 0.3252 - precision_m: 0.7086 - recall_m: 0.0391 - val_loss: 1.7277 - val_acc: 0.2745 - val_precision_m: 0.7722 - val_recall_m: 0.0300\n",
      "Epoch 40/1000\n",
      " - 2s - loss: 1.6791 - acc: 0.3366 - precision_m: 0.7156 - recall_m: 0.0402 - val_loss: 1.7189 - val_acc: 0.2789 - val_precision_m: 0.7222 - val_recall_m: 0.0311\n",
      "Epoch 41/1000\n",
      " - 3s - loss: 1.6742 - acc: 0.3300 - precision_m: 0.7422 - recall_m: 0.0419 - val_loss: 1.7119 - val_acc: 0.2856 - val_precision_m: 0.7556 - val_recall_m: 0.0331\n",
      "Epoch 42/1000\n",
      " - 3s - loss: 1.6640 - acc: 0.3381 - precision_m: 0.7318 - recall_m: 0.0464 - val_loss: 1.7053 - val_acc: 0.2922 - val_precision_m: 0.7444 - val_recall_m: 0.0321\n",
      "Epoch 43/1000\n",
      " - 3s - loss: 1.6545 - acc: 0.3445 - precision_m: 0.7245 - recall_m: 0.0481 - val_loss: 1.7033 - val_acc: 0.3109 - val_precision_m: 0.7844 - val_recall_m: 0.0444\n",
      "Epoch 44/1000\n",
      " - 3s - loss: 1.6477 - acc: 0.3434 - precision_m: 0.7063 - recall_m: 0.0507 - val_loss: 1.6935 - val_acc: 0.3021 - val_precision_m: 0.7822 - val_recall_m: 0.0373\n",
      "Epoch 45/1000\n",
      " - 3s - loss: 1.6367 - acc: 0.3511 - precision_m: 0.7301 - recall_m: 0.0551 - val_loss: 1.6888 - val_acc: 0.3109 - val_precision_m: 0.7644 - val_recall_m: 0.0446\n",
      "Epoch 46/1000\n",
      " - 3s - loss: 1.6309 - acc: 0.3527 - precision_m: 0.7256 - recall_m: 0.0575 - val_loss: 1.6846 - val_acc: 0.2966 - val_precision_m: 0.7733 - val_recall_m: 0.0475\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 47/1000\n",
      " - 3s - loss: 1.6263 - acc: 0.3556 - precision_m: 0.6964 - recall_m: 0.0581 - val_loss: 1.6762 - val_acc: 0.3153 - val_precision_m: 0.7756 - val_recall_m: 0.0496\n",
      "Epoch 48/1000\n",
      " - 3s - loss: 1.6189 - acc: 0.3561 - precision_m: 0.7260 - recall_m: 0.0635 - val_loss: 1.6734 - val_acc: 0.3098 - val_precision_m: 0.7533 - val_recall_m: 0.0548\n",
      "Epoch 49/1000\n",
      " - 3s - loss: 1.6131 - acc: 0.3681 - precision_m: 0.7046 - recall_m: 0.0625 - val_loss: 1.6656 - val_acc: 0.3208 - val_precision_m: 0.7649 - val_recall_m: 0.0548\n",
      "Epoch 50/1000\n",
      " - 3s - loss: 1.6028 - acc: 0.3637 - precision_m: 0.7171 - recall_m: 0.0669 - val_loss: 1.6595 - val_acc: 0.3065 - val_precision_m: 0.7710 - val_recall_m: 0.0580\n",
      "Epoch 51/1000\n",
      " - 3s - loss: 1.5963 - acc: 0.3763 - precision_m: 0.7124 - recall_m: 0.0724 - val_loss: 1.6545 - val_acc: 0.3252 - val_precision_m: 0.7660 - val_recall_m: 0.0559\n",
      "Epoch 52/1000\n",
      " - 3s - loss: 1.5859 - acc: 0.3777 - precision_m: 0.7075 - recall_m: 0.0752 - val_loss: 1.6500 - val_acc: 0.3241 - val_precision_m: 0.7571 - val_recall_m: 0.0611\n",
      "Epoch 53/1000\n",
      " - 3s - loss: 1.5811 - acc: 0.3817 - precision_m: 0.7252 - recall_m: 0.0798 - val_loss: 1.6460 - val_acc: 0.3352 - val_precision_m: 0.7408 - val_recall_m: 0.0611\n",
      "Epoch 54/1000\n",
      " - 3s - loss: 1.5722 - acc: 0.3826 - precision_m: 0.7194 - recall_m: 0.0792 - val_loss: 1.6404 - val_acc: 0.3286 - val_precision_m: 0.7417 - val_recall_m: 0.0705\n",
      "Epoch 55/1000\n",
      " - 3s - loss: 1.5668 - acc: 0.3922 - precision_m: 0.7102 - recall_m: 0.0825 - val_loss: 1.6349 - val_acc: 0.3341 - val_precision_m: 0.7737 - val_recall_m: 0.0705\n",
      "Epoch 56/1000\n",
      " - 3s - loss: 1.5600 - acc: 0.3907 - precision_m: 0.7185 - recall_m: 0.0900 - val_loss: 1.6256 - val_acc: 0.3385 - val_precision_m: 0.7994 - val_recall_m: 0.0746\n",
      "Epoch 57/1000\n",
      " - 3s - loss: 1.5463 - acc: 0.3939 - precision_m: 0.7097 - recall_m: 0.0909 - val_loss: 1.6188 - val_acc: 0.3363 - val_precision_m: 0.7624 - val_recall_m: 0.0715\n",
      "Epoch 58/1000\n",
      " - 3s - loss: 1.5437 - acc: 0.4006 - precision_m: 0.7112 - recall_m: 0.0972 - val_loss: 1.6117 - val_acc: 0.3407 - val_precision_m: 0.7832 - val_recall_m: 0.0798\n",
      "Epoch 59/1000\n",
      " - 3s - loss: 1.5357 - acc: 0.4033 - precision_m: 0.7348 - recall_m: 0.1000 - val_loss: 1.6055 - val_acc: 0.3451 - val_precision_m: 0.8151 - val_recall_m: 0.0850\n",
      "Epoch 60/1000\n",
      " - 3s - loss: 1.5275 - acc: 0.4065 - precision_m: 0.7164 - recall_m: 0.1021 - val_loss: 1.6014 - val_acc: 0.3407 - val_precision_m: 0.7858 - val_recall_m: 0.0840\n",
      "Epoch 61/1000\n",
      " - 3s - loss: 1.5163 - acc: 0.4143 - precision_m: 0.7066 - recall_m: 0.1080 - val_loss: 1.5991 - val_acc: 0.3484 - val_precision_m: 0.7566 - val_recall_m: 0.0850\n",
      "Epoch 62/1000\n",
      " - 3s - loss: 1.5134 - acc: 0.4168 - precision_m: 0.7167 - recall_m: 0.1090 - val_loss: 1.5906 - val_acc: 0.3418 - val_precision_m: 0.7735 - val_recall_m: 0.0840\n",
      "Epoch 63/1000\n",
      " - 3s - loss: 1.5069 - acc: 0.4194 - precision_m: 0.7084 - recall_m: 0.1115 - val_loss: 1.5857 - val_acc: 0.3418 - val_precision_m: 0.7819 - val_recall_m: 0.0902\n",
      "Epoch 64/1000\n",
      " - 3s - loss: 1.4986 - acc: 0.4143 - precision_m: 0.7221 - recall_m: 0.1169 - val_loss: 1.5787 - val_acc: 0.3605 - val_precision_m: 0.7769 - val_recall_m: 0.0902\n",
      "Epoch 65/1000\n",
      " - 3s - loss: 1.4915 - acc: 0.4254 - precision_m: 0.7285 - recall_m: 0.1191 - val_loss: 1.5726 - val_acc: 0.3605 - val_precision_m: 0.7450 - val_recall_m: 0.0892\n",
      "Epoch 66/1000\n",
      " - 3s - loss: 1.4824 - acc: 0.4287 - precision_m: 0.7256 - recall_m: 0.1283 - val_loss: 1.5715 - val_acc: 0.3539 - val_precision_m: 0.7848 - val_recall_m: 0.0975\n",
      "Epoch 67/1000\n",
      " - 3s - loss: 1.4803 - acc: 0.4305 - precision_m: 0.7194 - recall_m: 0.1245 - val_loss: 1.5628 - val_acc: 0.3583 - val_precision_m: 0.7853 - val_recall_m: 0.1007\n",
      "Epoch 68/1000\n",
      " - 3s - loss: 1.4681 - acc: 0.4425 - precision_m: 0.7159 - recall_m: 0.1332 - val_loss: 1.5559 - val_acc: 0.3682 - val_precision_m: 0.7592 - val_recall_m: 0.0975\n",
      "Epoch 69/1000\n",
      " - 3s - loss: 1.4571 - acc: 0.4335 - precision_m: 0.7191 - recall_m: 0.1381 - val_loss: 1.5468 - val_acc: 0.3848 - val_precision_m: 0.7374 - val_recall_m: 0.1059\n",
      "Epoch 70/1000\n",
      " - 3s - loss: 1.4548 - acc: 0.4424 - precision_m: 0.7213 - recall_m: 0.1404 - val_loss: 1.5457 - val_acc: 0.3649 - val_precision_m: 0.7576 - val_recall_m: 0.1111\n",
      "Epoch 71/1000\n",
      " - 3s - loss: 1.4426 - acc: 0.4456 - precision_m: 0.7215 - recall_m: 0.1465 - val_loss: 1.5402 - val_acc: 0.3705 - val_precision_m: 0.7317 - val_recall_m: 0.1048\n",
      "Epoch 72/1000\n",
      " - 3s - loss: 1.4347 - acc: 0.4521 - precision_m: 0.7325 - recall_m: 0.1519 - val_loss: 1.5309 - val_acc: 0.3727 - val_precision_m: 0.7693 - val_recall_m: 0.1132\n",
      "Epoch 73/1000\n",
      " - 3s - loss: 1.4298 - acc: 0.4563 - precision_m: 0.7269 - recall_m: 0.1588 - val_loss: 1.5256 - val_acc: 0.3693 - val_precision_m: 0.7659 - val_recall_m: 0.1173\n",
      "Epoch 74/1000\n",
      " - 3s - loss: 1.4247 - acc: 0.4600 - precision_m: 0.7170 - recall_m: 0.1559 - val_loss: 1.5230 - val_acc: 0.3660 - val_precision_m: 0.7234 - val_recall_m: 0.1265\n",
      "Epoch 75/1000\n",
      " - 3s - loss: 1.4159 - acc: 0.4592 - precision_m: 0.7149 - recall_m: 0.1646 - val_loss: 1.5193 - val_acc: 0.3760 - val_precision_m: 0.7130 - val_recall_m: 0.1215\n",
      "Epoch 76/1000\n",
      " - 3s - loss: 1.4044 - acc: 0.4662 - precision_m: 0.7279 - recall_m: 0.1744 - val_loss: 1.5035 - val_acc: 0.3870 - val_precision_m: 0.7740 - val_recall_m: 0.1267\n",
      "Epoch 77/1000\n",
      " - 3s - loss: 1.3973 - acc: 0.4713 - precision_m: 0.7369 - recall_m: 0.1785 - val_loss: 1.4993 - val_acc: 0.3903 - val_precision_m: 0.7604 - val_recall_m: 0.1288\n",
      "Epoch 78/1000\n",
      " - 3s - loss: 1.3872 - acc: 0.4761 - precision_m: 0.7263 - recall_m: 0.1848 - val_loss: 1.4897 - val_acc: 0.3936 - val_precision_m: 0.7488 - val_recall_m: 0.1257\n",
      "Epoch 79/1000\n",
      " - 3s - loss: 1.3844 - acc: 0.4745 - precision_m: 0.7217 - recall_m: 0.1813 - val_loss: 1.4867 - val_acc: 0.3936 - val_precision_m: 0.7584 - val_recall_m: 0.1288\n",
      "Epoch 80/1000\n",
      " - 3s - loss: 1.3747 - acc: 0.4812 - precision_m: 0.7420 - recall_m: 0.1921 - val_loss: 1.4770 - val_acc: 0.3980 - val_precision_m: 0.7523 - val_recall_m: 0.1382\n",
      "Epoch 81/1000\n",
      " - 3s - loss: 1.3625 - acc: 0.4806 - precision_m: 0.7280 - recall_m: 0.2000 - val_loss: 1.4711 - val_acc: 0.3991 - val_precision_m: 0.7440 - val_recall_m: 0.1382\n",
      "Epoch 82/1000\n",
      " - 3s - loss: 1.3599 - acc: 0.4885 - precision_m: 0.7272 - recall_m: 0.1998 - val_loss: 1.4631 - val_acc: 0.4002 - val_precision_m: 0.7257 - val_recall_m: 0.1465\n",
      "Epoch 83/1000\n",
      " - 3s - loss: 1.3485 - acc: 0.4945 - precision_m: 0.7406 - recall_m: 0.2065 - val_loss: 1.4576 - val_acc: 0.4024 - val_precision_m: 0.7365 - val_recall_m: 0.1455\n",
      "Epoch 84/1000\n",
      " - 3s - loss: 1.3437 - acc: 0.4956 - precision_m: 0.7266 - recall_m: 0.2091 - val_loss: 1.4503 - val_acc: 0.4135 - val_precision_m: 0.7548 - val_recall_m: 0.1548\n",
      "Epoch 85/1000\n",
      " - 3s - loss: 1.3331 - acc: 0.4947 - precision_m: 0.7308 - recall_m: 0.2161 - val_loss: 1.4463 - val_acc: 0.4190 - val_precision_m: 0.7291 - val_recall_m: 0.1692\n",
      "Epoch 86/1000\n",
      " - 3s - loss: 1.3209 - acc: 0.5049 - precision_m: 0.7409 - recall_m: 0.2271 - val_loss: 1.4446 - val_acc: 0.4223 - val_precision_m: 0.7329 - val_recall_m: 0.1538\n",
      "Epoch 87/1000\n",
      " - 3s - loss: 1.3161 - acc: 0.5077 - precision_m: 0.7379 - recall_m: 0.2277 - val_loss: 1.4317 - val_acc: 0.4223 - val_precision_m: 0.7282 - val_recall_m: 0.1569\n",
      "Epoch 88/1000\n",
      " - 3s - loss: 1.3098 - acc: 0.5045 - precision_m: 0.7401 - recall_m: 0.2324 - val_loss: 1.4267 - val_acc: 0.4300 - val_precision_m: 0.7205 - val_recall_m: 0.1621\n",
      "Epoch 89/1000\n",
      " - 3s - loss: 1.3036 - acc: 0.5123 - precision_m: 0.7414 - recall_m: 0.2421 - val_loss: 1.4167 - val_acc: 0.4245 - val_precision_m: 0.7411 - val_recall_m: 0.1722\n",
      "Epoch 90/1000\n",
      " - 3s - loss: 1.2905 - acc: 0.5199 - precision_m: 0.7484 - recall_m: 0.2440 - val_loss: 1.4114 - val_acc: 0.4377 - val_precision_m: 0.6962 - val_recall_m: 0.1744\n",
      "Epoch 91/1000\n",
      " - 3s - loss: 1.2787 - acc: 0.5243 - precision_m: 0.7391 - recall_m: 0.2487 - val_loss: 1.4002 - val_acc: 0.4421 - val_precision_m: 0.7077 - val_recall_m: 0.1857\n",
      "Epoch 92/1000\n",
      " - 3s - loss: 1.2695 - acc: 0.5194 - precision_m: 0.7446 - recall_m: 0.2568 - val_loss: 1.3969 - val_acc: 0.4410 - val_precision_m: 0.7012 - val_recall_m: 0.1878\n",
      "Epoch 93/1000\n",
      " - 3s - loss: 1.2689 - acc: 0.5254 - precision_m: 0.7399 - recall_m: 0.2598 - val_loss: 1.3850 - val_acc: 0.4520 - val_precision_m: 0.7164 - val_recall_m: 0.1859\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 94/1000\n",
      " - 3s - loss: 1.2661 - acc: 0.5309 - precision_m: 0.7487 - recall_m: 0.2644 - val_loss: 1.3767 - val_acc: 0.4653 - val_precision_m: 0.7180 - val_recall_m: 0.1840\n",
      "Epoch 95/1000\n",
      " - 3s - loss: 1.2418 - acc: 0.5425 - precision_m: 0.7567 - recall_m: 0.2772 - val_loss: 1.3639 - val_acc: 0.4708 - val_precision_m: 0.7293 - val_recall_m: 0.2045\n",
      "Epoch 96/1000\n",
      " - 3s - loss: 1.2352 - acc: 0.5400 - precision_m: 0.7551 - recall_m: 0.2765 - val_loss: 1.3584 - val_acc: 0.4697 - val_precision_m: 0.7258 - val_recall_m: 0.2005\n",
      "Epoch 97/1000\n",
      " - 3s - loss: 1.2272 - acc: 0.5488 - precision_m: 0.7561 - recall_m: 0.2876 - val_loss: 1.3575 - val_acc: 0.4620 - val_precision_m: 0.7153 - val_recall_m: 0.2045\n",
      "Epoch 98/1000\n",
      " - 3s - loss: 1.2259 - acc: 0.5515 - precision_m: 0.7439 - recall_m: 0.2853 - val_loss: 1.3417 - val_acc: 0.4851 - val_precision_m: 0.7278 - val_recall_m: 0.2117\n",
      "Epoch 99/1000\n",
      " - 3s - loss: 1.2126 - acc: 0.5545 - precision_m: 0.7587 - recall_m: 0.2939 - val_loss: 1.3372 - val_acc: 0.4708 - val_precision_m: 0.7226 - val_recall_m: 0.2295\n",
      "Epoch 100/1000\n",
      " - 3s - loss: 1.2004 - acc: 0.5570 - precision_m: 0.7564 - recall_m: 0.3044 - val_loss: 1.3270 - val_acc: 0.4818 - val_precision_m: 0.7389 - val_recall_m: 0.2284\n",
      "Epoch 101/1000\n",
      " - 3s - loss: 1.1891 - acc: 0.5626 - precision_m: 0.7666 - recall_m: 0.3123 - val_loss: 1.3188 - val_acc: 0.4950 - val_precision_m: 0.7352 - val_recall_m: 0.2326\n",
      "Epoch 102/1000\n",
      " - 3s - loss: 1.1925 - acc: 0.5553 - precision_m: 0.7500 - recall_m: 0.3119 - val_loss: 1.3101 - val_acc: 0.4983 - val_precision_m: 0.7493 - val_recall_m: 0.2451\n",
      "Epoch 103/1000\n",
      " - 3s - loss: 1.1770 - acc: 0.5741 - precision_m: 0.7648 - recall_m: 0.3210 - val_loss: 1.3027 - val_acc: 0.4994 - val_precision_m: 0.7485 - val_recall_m: 0.2513\n",
      "Epoch 104/1000\n",
      " - 3s - loss: 1.1684 - acc: 0.5701 - precision_m: 0.7688 - recall_m: 0.3298 - val_loss: 1.2938 - val_acc: 0.5094 - val_precision_m: 0.7729 - val_recall_m: 0.2607\n",
      "Epoch 105/1000\n",
      " - 3s - loss: 1.1538 - acc: 0.5746 - precision_m: 0.7691 - recall_m: 0.3356 - val_loss: 1.2800 - val_acc: 0.5182 - val_precision_m: 0.7688 - val_recall_m: 0.2524\n",
      "Epoch 106/1000\n",
      " - 3s - loss: 1.1448 - acc: 0.5848 - precision_m: 0.7689 - recall_m: 0.3412 - val_loss: 1.2755 - val_acc: 0.5094 - val_precision_m: 0.7559 - val_recall_m: 0.2670\n",
      "Epoch 107/1000\n",
      " - 3s - loss: 1.1484 - acc: 0.5806 - precision_m: 0.7577 - recall_m: 0.3386 - val_loss: 1.2666 - val_acc: 0.5237 - val_precision_m: 0.7744 - val_recall_m: 0.2722\n",
      "Epoch 108/1000\n",
      " - 3s - loss: 1.1329 - acc: 0.5869 - precision_m: 0.7686 - recall_m: 0.3544 - val_loss: 1.2592 - val_acc: 0.5303 - val_precision_m: 0.7846 - val_recall_m: 0.2753\n",
      "Epoch 109/1000\n",
      " - 3s - loss: 1.1179 - acc: 0.5955 - precision_m: 0.7804 - recall_m: 0.3627 - val_loss: 1.2485 - val_acc: 0.5259 - val_precision_m: 0.7844 - val_recall_m: 0.2847\n",
      "Epoch 110/1000\n",
      " - 3s - loss: 1.1083 - acc: 0.5975 - precision_m: 0.7837 - recall_m: 0.3693 - val_loss: 1.2424 - val_acc: 0.5380 - val_precision_m: 0.7723 - val_recall_m: 0.2899\n",
      "Epoch 111/1000\n",
      " - 3s - loss: 1.1115 - acc: 0.5973 - precision_m: 0.7753 - recall_m: 0.3717 - val_loss: 1.2329 - val_acc: 0.5358 - val_precision_m: 0.7790 - val_recall_m: 0.2940\n",
      "Epoch 112/1000\n",
      " - 3s - loss: 1.0990 - acc: 0.5945 - precision_m: 0.7716 - recall_m: 0.3713 - val_loss: 1.2278 - val_acc: 0.5424 - val_precision_m: 0.7864 - val_recall_m: 0.2940\n",
      "Epoch 113/1000\n",
      " - 3s - loss: 1.0830 - acc: 0.6086 - precision_m: 0.7763 - recall_m: 0.3884 - val_loss: 1.2138 - val_acc: 0.5579 - val_precision_m: 0.7917 - val_recall_m: 0.3086\n",
      "Epoch 114/1000\n",
      " - 3s - loss: 1.0771 - acc: 0.6113 - precision_m: 0.7812 - recall_m: 0.3895 - val_loss: 1.2153 - val_acc: 0.5535 - val_precision_m: 0.7718 - val_recall_m: 0.3107\n",
      "Epoch 115/1000\n",
      " - 3s - loss: 1.0633 - acc: 0.6140 - precision_m: 0.7796 - recall_m: 0.3979 - val_loss: 1.1996 - val_acc: 0.5645 - val_precision_m: 0.7957 - val_recall_m: 0.3211\n",
      "Epoch 116/1000\n",
      " - 3s - loss: 1.0584 - acc: 0.6181 - precision_m: 0.7811 - recall_m: 0.4015 - val_loss: 1.1936 - val_acc: 0.5634 - val_precision_m: 0.7897 - val_recall_m: 0.3190\n",
      "Epoch 117/1000\n",
      " - 3s - loss: 1.0557 - acc: 0.6154 - precision_m: 0.7805 - recall_m: 0.4053 - val_loss: 1.1852 - val_acc: 0.5656 - val_precision_m: 0.8002 - val_recall_m: 0.3190\n",
      "Epoch 118/1000\n",
      " - 3s - loss: 1.0481 - acc: 0.6241 - precision_m: 0.7778 - recall_m: 0.4033 - val_loss: 1.1710 - val_acc: 0.5601 - val_precision_m: 0.7912 - val_recall_m: 0.3347\n",
      "Epoch 119/1000\n",
      " - 3s - loss: 1.0306 - acc: 0.6246 - precision_m: 0.7922 - recall_m: 0.4144 - val_loss: 1.1593 - val_acc: 0.5755 - val_precision_m: 0.7994 - val_recall_m: 0.3420\n",
      "Epoch 120/1000\n",
      " - 3s - loss: 1.0230 - acc: 0.6315 - precision_m: 0.7952 - recall_m: 0.4279 - val_loss: 1.1545 - val_acc: 0.5821 - val_precision_m: 0.7819 - val_recall_m: 0.3409\n",
      "Epoch 121/1000\n",
      " - 3s - loss: 1.0076 - acc: 0.6401 - precision_m: 0.8002 - recall_m: 0.4357 - val_loss: 1.1449 - val_acc: 0.5865 - val_precision_m: 0.8015 - val_recall_m: 0.3451\n",
      "Epoch 122/1000\n",
      " - 3s - loss: 0.9990 - acc: 0.6453 - precision_m: 0.7968 - recall_m: 0.4402 - val_loss: 1.1365 - val_acc: 0.5888 - val_precision_m: 0.7928 - val_recall_m: 0.3545\n",
      "Epoch 123/1000\n",
      " - 3s - loss: 0.9935 - acc: 0.6467 - precision_m: 0.8009 - recall_m: 0.4406 - val_loss: 1.1237 - val_acc: 0.5998 - val_precision_m: 0.7880 - val_recall_m: 0.3701\n",
      "Epoch 124/1000\n",
      " - 3s - loss: 0.9895 - acc: 0.6480 - precision_m: 0.7967 - recall_m: 0.4507 - val_loss: 1.1217 - val_acc: 0.5943 - val_precision_m: 0.7902 - val_recall_m: 0.3638\n",
      "Epoch 125/1000\n",
      " - 3s - loss: 0.9799 - acc: 0.6483 - precision_m: 0.7945 - recall_m: 0.4504 - val_loss: 1.1107 - val_acc: 0.6053 - val_precision_m: 0.7968 - val_recall_m: 0.3690\n",
      "Epoch 126/1000\n",
      " - 3s - loss: 0.9641 - acc: 0.6624 - precision_m: 0.8115 - recall_m: 0.4635 - val_loss: 1.0969 - val_acc: 0.6075 - val_precision_m: 0.7963 - val_recall_m: 0.3784\n",
      "Epoch 127/1000\n",
      " - 3s - loss: 0.9592 - acc: 0.6600 - precision_m: 0.8063 - recall_m: 0.4731 - val_loss: 1.0934 - val_acc: 0.6141 - val_precision_m: 0.8018 - val_recall_m: 0.3815\n",
      "Epoch 128/1000\n",
      " - 3s - loss: 0.9415 - acc: 0.6668 - precision_m: 0.8078 - recall_m: 0.4792 - val_loss: 1.0802 - val_acc: 0.6196 - val_precision_m: 0.7933 - val_recall_m: 0.3857\n",
      "Epoch 129/1000\n",
      " - 3s - loss: 0.9398 - acc: 0.6631 - precision_m: 0.8103 - recall_m: 0.4833 - val_loss: 1.0753 - val_acc: 0.6174 - val_precision_m: 0.8018 - val_recall_m: 0.3959\n",
      "Epoch 130/1000\n",
      " - 3s - loss: 0.9285 - acc: 0.6712 - precision_m: 0.8157 - recall_m: 0.4948 - val_loss: 1.0626 - val_acc: 0.6196 - val_precision_m: 0.7957 - val_recall_m: 0.4053\n",
      "Epoch 131/1000\n",
      " - 3s - loss: 0.9168 - acc: 0.6743 - precision_m: 0.8100 - recall_m: 0.4944 - val_loss: 1.0553 - val_acc: 0.6329 - val_precision_m: 0.8024 - val_recall_m: 0.3982\n",
      "Epoch 132/1000\n",
      " - 3s - loss: 0.9113 - acc: 0.6687 - precision_m: 0.8192 - recall_m: 0.5012 - val_loss: 1.0549 - val_acc: 0.6185 - val_precision_m: 0.8030 - val_recall_m: 0.4084\n",
      "Epoch 133/1000\n",
      " - 3s - loss: 0.9137 - acc: 0.6724 - precision_m: 0.8161 - recall_m: 0.4982 - val_loss: 1.0460 - val_acc: 0.6417 - val_precision_m: 0.8114 - val_recall_m: 0.4149\n",
      "Epoch 134/1000\n",
      " - 3s - loss: 0.8946 - acc: 0.6841 - precision_m: 0.8194 - recall_m: 0.5110 - val_loss: 1.0246 - val_acc: 0.6373 - val_precision_m: 0.8220 - val_recall_m: 0.4272\n",
      "Epoch 135/1000\n",
      " - 3s - loss: 0.8762 - acc: 0.6974 - precision_m: 0.8231 - recall_m: 0.5223 - val_loss: 1.0139 - val_acc: 0.6483 - val_precision_m: 0.8200 - val_recall_m: 0.4355\n",
      "Epoch 136/1000\n",
      " - 3s - loss: 0.8717 - acc: 0.6914 - precision_m: 0.8149 - recall_m: 0.5266 - val_loss: 1.0041 - val_acc: 0.6527 - val_precision_m: 0.8187 - val_recall_m: 0.4376\n",
      "Epoch 137/1000\n",
      " - 3s - loss: 0.8679 - acc: 0.6951 - precision_m: 0.8213 - recall_m: 0.5328 - val_loss: 0.9907 - val_acc: 0.6483 - val_precision_m: 0.8212 - val_recall_m: 0.4376\n",
      "Epoch 138/1000\n",
      " - 3s - loss: 0.8662 - acc: 0.6929 - precision_m: 0.8210 - recall_m: 0.5327 - val_loss: 0.9861 - val_acc: 0.6593 - val_precision_m: 0.8285 - val_recall_m: 0.4501\n",
      "Epoch 139/1000\n",
      " - 3s - loss: 0.8586 - acc: 0.6945 - precision_m: 0.8195 - recall_m: 0.5360 - val_loss: 0.9865 - val_acc: 0.6648 - val_precision_m: 0.8261 - val_recall_m: 0.4668\n",
      "Epoch 140/1000\n",
      " - 3s - loss: 0.8343 - acc: 0.7066 - precision_m: 0.8280 - recall_m: 0.5481 - val_loss: 0.9712 - val_acc: 0.6571 - val_precision_m: 0.8138 - val_recall_m: 0.4697\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 141/1000\n",
      " - 3s - loss: 0.8287 - acc: 0.7071 - precision_m: 0.8300 - recall_m: 0.5544 - val_loss: 0.9579 - val_acc: 0.6692 - val_precision_m: 0.8394 - val_recall_m: 0.4688\n",
      "Epoch 142/1000\n",
      " - 3s - loss: 0.8236 - acc: 0.7092 - precision_m: 0.8301 - recall_m: 0.5583 - val_loss: 0.9457 - val_acc: 0.6714 - val_precision_m: 0.8363 - val_recall_m: 0.4895\n",
      "Epoch 143/1000\n",
      " - 3s - loss: 0.8153 - acc: 0.7135 - precision_m: 0.8309 - recall_m: 0.5656 - val_loss: 0.9471 - val_acc: 0.6736 - val_precision_m: 0.8263 - val_recall_m: 0.4876\n",
      "Epoch 144/1000\n",
      " - 3s - loss: 0.7971 - acc: 0.7229 - precision_m: 0.8408 - recall_m: 0.5768 - val_loss: 0.9365 - val_acc: 0.6770 - val_precision_m: 0.8392 - val_recall_m: 0.4938\n",
      "Epoch 145/1000\n",
      " - 3s - loss: 0.8028 - acc: 0.7167 - precision_m: 0.8282 - recall_m: 0.5768 - val_loss: 0.9342 - val_acc: 0.6759 - val_precision_m: 0.8350 - val_recall_m: 0.5009\n",
      "Epoch 146/1000\n",
      " - 3s - loss: 0.8041 - acc: 0.7200 - precision_m: 0.8290 - recall_m: 0.5735 - val_loss: 0.9191 - val_acc: 0.6880 - val_precision_m: 0.8384 - val_recall_m: 0.5011\n",
      "Epoch 147/1000\n",
      " - 3s - loss: 0.7918 - acc: 0.7237 - precision_m: 0.8372 - recall_m: 0.5846 - val_loss: 0.9060 - val_acc: 0.6847 - val_precision_m: 0.8394 - val_recall_m: 0.5207\n",
      "Epoch 148/1000\n",
      " - 3s - loss: 0.7753 - acc: 0.7289 - precision_m: 0.8424 - recall_m: 0.5964 - val_loss: 0.8932 - val_acc: 0.6946 - val_precision_m: 0.8548 - val_recall_m: 0.5197\n",
      "Epoch 149/1000\n",
      " - 3s - loss: 0.7710 - acc: 0.7350 - precision_m: 0.8357 - recall_m: 0.5974 - val_loss: 0.8953 - val_acc: 0.6913 - val_precision_m: 0.8516 - val_recall_m: 0.5166\n",
      "Epoch 150/1000\n",
      " - 3s - loss: 0.7541 - acc: 0.7374 - precision_m: 0.8412 - recall_m: 0.6073 - val_loss: 0.8776 - val_acc: 0.7111 - val_precision_m: 0.8543 - val_recall_m: 0.5332\n",
      "Epoch 151/1000\n",
      " - 3s - loss: 0.7557 - acc: 0.7315 - precision_m: 0.8388 - recall_m: 0.5973 - val_loss: 0.8644 - val_acc: 0.7111 - val_precision_m: 0.8597 - val_recall_m: 0.5437\n",
      "Epoch 152/1000\n",
      " - 3s - loss: 0.7419 - acc: 0.7371 - precision_m: 0.8439 - recall_m: 0.6079 - val_loss: 0.8528 - val_acc: 0.7111 - val_precision_m: 0.8551 - val_recall_m: 0.5447\n",
      "Epoch 153/1000\n",
      " - 3s - loss: 0.7305 - acc: 0.7483 - precision_m: 0.8489 - recall_m: 0.6191 - val_loss: 0.8466 - val_acc: 0.7133 - val_precision_m: 0.8607 - val_recall_m: 0.5572\n",
      "Epoch 154/1000\n",
      " - 3s - loss: 0.7141 - acc: 0.7517 - precision_m: 0.8503 - recall_m: 0.6321 - val_loss: 0.8378 - val_acc: 0.7222 - val_precision_m: 0.8607 - val_recall_m: 0.5572\n",
      "Epoch 155/1000\n",
      " - 3s - loss: 0.7222 - acc: 0.7484 - precision_m: 0.8480 - recall_m: 0.6228 - val_loss: 0.8311 - val_acc: 0.7277 - val_precision_m: 0.8686 - val_recall_m: 0.5707\n",
      "Epoch 156/1000\n",
      " - 3s - loss: 0.7023 - acc: 0.7567 - precision_m: 0.8510 - recall_m: 0.6432 - val_loss: 0.8201 - val_acc: 0.7310 - val_precision_m: 0.8692 - val_recall_m: 0.5707\n",
      "Epoch 157/1000\n",
      " - 3s - loss: 0.6938 - acc: 0.7603 - precision_m: 0.8514 - recall_m: 0.6407 - val_loss: 0.8145 - val_acc: 0.7365 - val_precision_m: 0.8579 - val_recall_m: 0.5780\n",
      "Epoch 158/1000\n",
      " - 3s - loss: 0.7024 - acc: 0.7623 - precision_m: 0.8519 - recall_m: 0.6434 - val_loss: 0.7951 - val_acc: 0.7343 - val_precision_m: 0.8779 - val_recall_m: 0.5916\n",
      "Epoch 159/1000\n",
      " - 3s - loss: 0.6927 - acc: 0.7593 - precision_m: 0.8513 - recall_m: 0.6436 - val_loss: 0.8046 - val_acc: 0.7420 - val_precision_m: 0.8786 - val_recall_m: 0.5770\n",
      "Epoch 160/1000\n",
      " - 3s - loss: 0.6833 - acc: 0.7599 - precision_m: 0.8504 - recall_m: 0.6459 - val_loss: 0.7840 - val_acc: 0.7453 - val_precision_m: 0.8763 - val_recall_m: 0.5926\n",
      "Epoch 161/1000\n",
      " - 3s - loss: 0.6584 - acc: 0.7805 - precision_m: 0.8637 - recall_m: 0.6675 - val_loss: 0.7853 - val_acc: 0.7409 - val_precision_m: 0.8676 - val_recall_m: 0.5884\n",
      "Epoch 162/1000\n",
      " - 3s - loss: 0.6646 - acc: 0.7752 - precision_m: 0.8607 - recall_m: 0.6616 - val_loss: 0.7670 - val_acc: 0.7497 - val_precision_m: 0.8786 - val_recall_m: 0.5999\n",
      "Epoch 163/1000\n",
      " - 3s - loss: 0.6564 - acc: 0.7762 - precision_m: 0.8613 - recall_m: 0.6672 - val_loss: 0.7649 - val_acc: 0.7552 - val_precision_m: 0.8888 - val_recall_m: 0.6072\n",
      "Epoch 164/1000\n",
      " - 3s - loss: 0.6512 - acc: 0.7757 - precision_m: 0.8596 - recall_m: 0.6700 - val_loss: 0.7629 - val_acc: 0.7486 - val_precision_m: 0.8724 - val_recall_m: 0.6062\n",
      "Epoch 165/1000\n",
      " - 3s - loss: 0.6377 - acc: 0.7843 - precision_m: 0.8627 - recall_m: 0.6804 - val_loss: 0.7525 - val_acc: 0.7652 - val_precision_m: 0.8722 - val_recall_m: 0.6145\n",
      "Epoch 166/1000\n",
      " - 3s - loss: 0.6310 - acc: 0.7827 - precision_m: 0.8690 - recall_m: 0.6798 - val_loss: 0.7319 - val_acc: 0.7685 - val_precision_m: 0.8890 - val_recall_m: 0.6280\n",
      "Epoch 167/1000\n",
      " - 3s - loss: 0.6188 - acc: 0.7870 - precision_m: 0.8650 - recall_m: 0.6841 - val_loss: 0.7276 - val_acc: 0.7685 - val_precision_m: 0.8883 - val_recall_m: 0.6291\n",
      "Epoch 168/1000\n",
      " - 3s - loss: 0.6058 - acc: 0.7978 - precision_m: 0.8741 - recall_m: 0.7030 - val_loss: 0.7188 - val_acc: 0.7729 - val_precision_m: 0.8841 - val_recall_m: 0.6322\n",
      "Epoch 169/1000\n",
      " - 3s - loss: 0.6095 - acc: 0.7939 - precision_m: 0.8686 - recall_m: 0.7036 - val_loss: 0.7073 - val_acc: 0.7817 - val_precision_m: 0.8973 - val_recall_m: 0.6416\n",
      "Epoch 170/1000\n",
      " - 3s - loss: 0.6172 - acc: 0.7902 - precision_m: 0.8648 - recall_m: 0.6884 - val_loss: 0.7059 - val_acc: 0.7751 - val_precision_m: 0.8893 - val_recall_m: 0.6528\n",
      "Epoch 171/1000\n",
      " - 3s - loss: 0.5991 - acc: 0.7963 - precision_m: 0.8686 - recall_m: 0.7037 - val_loss: 0.6879 - val_acc: 0.7795 - val_precision_m: 0.8973 - val_recall_m: 0.6551\n",
      "Epoch 172/1000\n",
      " - 3s - loss: 0.5936 - acc: 0.7950 - precision_m: 0.8691 - recall_m: 0.7073 - val_loss: 0.6849 - val_acc: 0.7905 - val_precision_m: 0.9044 - val_recall_m: 0.6572\n",
      "Epoch 173/1000\n",
      " - 3s - loss: 0.5893 - acc: 0.7984 - precision_m: 0.8746 - recall_m: 0.7045 - val_loss: 0.6789 - val_acc: 0.7938 - val_precision_m: 0.8929 - val_recall_m: 0.6447\n",
      "Epoch 174/1000\n",
      " - 3s - loss: 0.5712 - acc: 0.8115 - precision_m: 0.8750 - recall_m: 0.7241 - val_loss: 0.6744 - val_acc: 0.7938 - val_precision_m: 0.8917 - val_recall_m: 0.6530\n",
      "Epoch 175/1000\n",
      " - 3s - loss: 0.5672 - acc: 0.8092 - precision_m: 0.8779 - recall_m: 0.7208 - val_loss: 0.6625 - val_acc: 0.7905 - val_precision_m: 0.8853 - val_recall_m: 0.6726\n",
      "Epoch 176/1000\n",
      " - 3s - loss: 0.5711 - acc: 0.8056 - precision_m: 0.8784 - recall_m: 0.7261 - val_loss: 0.6545 - val_acc: 0.7993 - val_precision_m: 0.9116 - val_recall_m: 0.6789\n",
      "Epoch 177/1000\n",
      " - 3s - loss: 0.5538 - acc: 0.8140 - precision_m: 0.8849 - recall_m: 0.7316 - val_loss: 0.6400 - val_acc: 0.8049 - val_precision_m: 0.9007 - val_recall_m: 0.6841\n",
      "Epoch 178/1000\n",
      " - 3s - loss: 0.5506 - acc: 0.8124 - precision_m: 0.8798 - recall_m: 0.7303 - val_loss: 0.6444 - val_acc: 0.8026 - val_precision_m: 0.9238 - val_recall_m: 0.6810\n",
      "Epoch 179/1000\n",
      " - 3s - loss: 0.5440 - acc: 0.8179 - precision_m: 0.8801 - recall_m: 0.7368 - val_loss: 0.6308 - val_acc: 0.8071 - val_precision_m: 0.9152 - val_recall_m: 0.6966\n",
      "Epoch 180/1000\n",
      " - 3s - loss: 0.5212 - acc: 0.8294 - precision_m: 0.8890 - recall_m: 0.7516 - val_loss: 0.6212 - val_acc: 0.8137 - val_precision_m: 0.9110 - val_recall_m: 0.7018\n",
      "Epoch 181/1000\n",
      " - 3s - loss: 0.5329 - acc: 0.8283 - precision_m: 0.8864 - recall_m: 0.7533 - val_loss: 0.6065 - val_acc: 0.8159 - val_precision_m: 0.9095 - val_recall_m: 0.7080\n",
      "Epoch 182/1000\n",
      " - 3s - loss: 0.5226 - acc: 0.8308 - precision_m: 0.8836 - recall_m: 0.7532 - val_loss: 0.5945 - val_acc: 0.8148 - val_precision_m: 0.9224 - val_recall_m: 0.7153\n",
      "Epoch 183/1000\n",
      " - 3s - loss: 0.5104 - acc: 0.8297 - precision_m: 0.8875 - recall_m: 0.7516 - val_loss: 0.5924 - val_acc: 0.8247 - val_precision_m: 0.9180 - val_recall_m: 0.7216\n",
      "Epoch 184/1000\n",
      " - 3s - loss: 0.5122 - acc: 0.8301 - precision_m: 0.8885 - recall_m: 0.7535 - val_loss: 0.5872 - val_acc: 0.8225 - val_precision_m: 0.9172 - val_recall_m: 0.7237\n",
      "Epoch 185/1000\n",
      " - 3s - loss: 0.5078 - acc: 0.8339 - precision_m: 0.8886 - recall_m: 0.7596 - val_loss: 0.5741 - val_acc: 0.8236 - val_precision_m: 0.9244 - val_recall_m: 0.7258\n",
      "Epoch 186/1000\n",
      " - 3s - loss: 0.4986 - acc: 0.8392 - precision_m: 0.8915 - recall_m: 0.7667 - val_loss: 0.5822 - val_acc: 0.8335 - val_precision_m: 0.9294 - val_recall_m: 0.7433\n",
      "Epoch 187/1000\n",
      " - 3s - loss: 0.4934 - acc: 0.8387 - precision_m: 0.8951 - recall_m: 0.7704 - val_loss: 0.5719 - val_acc: 0.8357 - val_precision_m: 0.9163 - val_recall_m: 0.7310\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 188/1000\n",
      " - 3s - loss: 0.4921 - acc: 0.8354 - precision_m: 0.8890 - recall_m: 0.7684 - val_loss: 0.5566 - val_acc: 0.8302 - val_precision_m: 0.9230 - val_recall_m: 0.7341\n",
      "Epoch 189/1000\n",
      " - 3s - loss: 0.4781 - acc: 0.8474 - precision_m: 0.8988 - recall_m: 0.7795 - val_loss: 0.5512 - val_acc: 0.8346 - val_precision_m: 0.9228 - val_recall_m: 0.7497\n",
      "Epoch 190/1000\n",
      " - 3s - loss: 0.4785 - acc: 0.8428 - precision_m: 0.8950 - recall_m: 0.7758 - val_loss: 0.5386 - val_acc: 0.8445 - val_precision_m: 0.9228 - val_recall_m: 0.7516\n",
      "Epoch 191/1000\n",
      " - 3s - loss: 0.4694 - acc: 0.8474 - precision_m: 0.8981 - recall_m: 0.7800 - val_loss: 0.5268 - val_acc: 0.8479 - val_precision_m: 0.9211 - val_recall_m: 0.7508\n",
      "Epoch 192/1000\n",
      " - 3s - loss: 0.4610 - acc: 0.8451 - precision_m: 0.8993 - recall_m: 0.7841 - val_loss: 0.5170 - val_acc: 0.8456 - val_precision_m: 0.9356 - val_recall_m: 0.7641\n",
      "Epoch 193/1000\n",
      " - 3s - loss: 0.4546 - acc: 0.8507 - precision_m: 0.8974 - recall_m: 0.7900 - val_loss: 0.5216 - val_acc: 0.8523 - val_precision_m: 0.9244 - val_recall_m: 0.7633\n",
      "Epoch 194/1000\n",
      " - 3s - loss: 0.4506 - acc: 0.8505 - precision_m: 0.8987 - recall_m: 0.7935 - val_loss: 0.5157 - val_acc: 0.8467 - val_precision_m: 0.9257 - val_recall_m: 0.7664\n",
      "Epoch 195/1000\n",
      " - 3s - loss: 0.4478 - acc: 0.8521 - precision_m: 0.9034 - recall_m: 0.7919 - val_loss: 0.5042 - val_acc: 0.8567 - val_precision_m: 0.9362 - val_recall_m: 0.7849\n",
      "Epoch 196/1000\n",
      " - 3s - loss: 0.4483 - acc: 0.8549 - precision_m: 0.8997 - recall_m: 0.7943 - val_loss: 0.4916 - val_acc: 0.8589 - val_precision_m: 0.9405 - val_recall_m: 0.7768\n",
      "Epoch 197/1000\n",
      " - 3s - loss: 0.4293 - acc: 0.8626 - precision_m: 0.9074 - recall_m: 0.8076 - val_loss: 0.4958 - val_acc: 0.8611 - val_precision_m: 0.9332 - val_recall_m: 0.7612\n",
      "Epoch 198/1000\n",
      " - 3s - loss: 0.4284 - acc: 0.8561 - precision_m: 0.9051 - recall_m: 0.8014 - val_loss: 0.4794 - val_acc: 0.8677 - val_precision_m: 0.9394 - val_recall_m: 0.7860\n",
      "Epoch 199/1000\n",
      " - 3s - loss: 0.4229 - acc: 0.8614 - precision_m: 0.9039 - recall_m: 0.8086 - val_loss: 0.4598 - val_acc: 0.8677 - val_precision_m: 0.9454 - val_recall_m: 0.7891\n",
      "Epoch 200/1000\n",
      " - 3s - loss: 0.4252 - acc: 0.8575 - precision_m: 0.9005 - recall_m: 0.8037 - val_loss: 0.4602 - val_acc: 0.8721 - val_precision_m: 0.9534 - val_recall_m: 0.7995\n",
      "Epoch 201/1000\n",
      " - 3s - loss: 0.4162 - acc: 0.8602 - precision_m: 0.9015 - recall_m: 0.8104 - val_loss: 0.4518 - val_acc: 0.8787 - val_precision_m: 0.9526 - val_recall_m: 0.8118\n",
      "Epoch 202/1000\n",
      " - 3s - loss: 0.4134 - acc: 0.8646 - precision_m: 0.9052 - recall_m: 0.8151 - val_loss: 0.4567 - val_acc: 0.8743 - val_precision_m: 0.9455 - val_recall_m: 0.7974\n",
      "Epoch 203/1000\n",
      " - 3s - loss: 0.4042 - acc: 0.8701 - precision_m: 0.9087 - recall_m: 0.8196 - val_loss: 0.4401 - val_acc: 0.8787 - val_precision_m: 0.9483 - val_recall_m: 0.8108\n",
      "Epoch 204/1000\n",
      " - 3s - loss: 0.3940 - acc: 0.8716 - precision_m: 0.9117 - recall_m: 0.8201 - val_loss: 0.4441 - val_acc: 0.8798 - val_precision_m: 0.9351 - val_recall_m: 0.8120\n",
      "Epoch 205/1000\n",
      " - 3s - loss: 0.4038 - acc: 0.8661 - precision_m: 0.9046 - recall_m: 0.8180 - val_loss: 0.4477 - val_acc: 0.8787 - val_precision_m: 0.9419 - val_recall_m: 0.8254\n",
      "Epoch 206/1000\n",
      " - 3s - loss: 0.3885 - acc: 0.8746 - precision_m: 0.9131 - recall_m: 0.8269 - val_loss: 0.4236 - val_acc: 0.8787 - val_precision_m: 0.9461 - val_recall_m: 0.8191\n",
      "Epoch 207/1000\n",
      " - 3s - loss: 0.3927 - acc: 0.8728 - precision_m: 0.9110 - recall_m: 0.8232 - val_loss: 0.4221 - val_acc: 0.8875 - val_precision_m: 0.9522 - val_recall_m: 0.8223\n",
      "Epoch 208/1000\n",
      " - 3s - loss: 0.3812 - acc: 0.8750 - precision_m: 0.9156 - recall_m: 0.8297 - val_loss: 0.4074 - val_acc: 0.8875 - val_precision_m: 0.9547 - val_recall_m: 0.8285\n",
      "Epoch 209/1000\n",
      " - 3s - loss: 0.3849 - acc: 0.8696 - precision_m: 0.9062 - recall_m: 0.8248 - val_loss: 0.3983 - val_acc: 0.8931 - val_precision_m: 0.9549 - val_recall_m: 0.8337\n",
      "Epoch 210/1000\n",
      " - 3s - loss: 0.3682 - acc: 0.8801 - precision_m: 0.9139 - recall_m: 0.8431 - val_loss: 0.3953 - val_acc: 0.8920 - val_precision_m: 0.9458 - val_recall_m: 0.8379\n",
      "Epoch 211/1000\n",
      " - 3s - loss: 0.3656 - acc: 0.8832 - precision_m: 0.9191 - recall_m: 0.8395 - val_loss: 0.3971 - val_acc: 0.8897 - val_precision_m: 0.9398 - val_recall_m: 0.8379\n",
      "Epoch 212/1000\n",
      " - 3s - loss: 0.3706 - acc: 0.8760 - precision_m: 0.9126 - recall_m: 0.8318 - val_loss: 0.3803 - val_acc: 0.9063 - val_precision_m: 0.9616 - val_recall_m: 0.8535\n",
      "Epoch 213/1000\n",
      " - 3s - loss: 0.3512 - acc: 0.8870 - precision_m: 0.9169 - recall_m: 0.8468 - val_loss: 0.3794 - val_acc: 0.8953 - val_precision_m: 0.9541 - val_recall_m: 0.8360\n",
      "Epoch 214/1000\n",
      " - 3s - loss: 0.3516 - acc: 0.8828 - precision_m: 0.9180 - recall_m: 0.8434 - val_loss: 0.3727 - val_acc: 0.9008 - val_precision_m: 0.9415 - val_recall_m: 0.8504\n",
      "Epoch 215/1000\n",
      " - 3s - loss: 0.3464 - acc: 0.8891 - precision_m: 0.9186 - recall_m: 0.8486 - val_loss: 0.3650 - val_acc: 0.9052 - val_precision_m: 0.9528 - val_recall_m: 0.8545\n",
      "Epoch 216/1000\n",
      " - 3s - loss: 0.3531 - acc: 0.8843 - precision_m: 0.9168 - recall_m: 0.8427 - val_loss: 0.3641 - val_acc: 0.9118 - val_precision_m: 0.9617 - val_recall_m: 0.8566\n",
      "Epoch 217/1000\n",
      " - 3s - loss: 0.3423 - acc: 0.8885 - precision_m: 0.9214 - recall_m: 0.8477 - val_loss: 0.3776 - val_acc: 0.9030 - val_precision_m: 0.9580 - val_recall_m: 0.8535\n",
      "Epoch 218/1000\n",
      " - 3s - loss: 0.3430 - acc: 0.8863 - precision_m: 0.9166 - recall_m: 0.8501 - val_loss: 0.3503 - val_acc: 0.9173 - val_precision_m: 0.9638 - val_recall_m: 0.8556\n",
      "Epoch 219/1000\n",
      " - 3s - loss: 0.3313 - acc: 0.8922 - precision_m: 0.9233 - recall_m: 0.8600 - val_loss: 0.3530 - val_acc: 0.9129 - val_precision_m: 0.9605 - val_recall_m: 0.8783\n",
      "Epoch 220/1000\n",
      " - 3s - loss: 0.3217 - acc: 0.8993 - precision_m: 0.9276 - recall_m: 0.8611 - val_loss: 0.3400 - val_acc: 0.9173 - val_precision_m: 0.9624 - val_recall_m: 0.8741\n",
      "Epoch 221/1000\n",
      " - 3s - loss: 0.3179 - acc: 0.8977 - precision_m: 0.9260 - recall_m: 0.8627 - val_loss: 0.3367 - val_acc: 0.9184 - val_precision_m: 0.9668 - val_recall_m: 0.8723\n",
      "Epoch 222/1000\n",
      " - 3s - loss: 0.3286 - acc: 0.8922 - precision_m: 0.9205 - recall_m: 0.8588 - val_loss: 0.3254 - val_acc: 0.9250 - val_precision_m: 0.9613 - val_recall_m: 0.8754\n",
      "Epoch 223/1000\n",
      " - 3s - loss: 0.3179 - acc: 0.8953 - precision_m: 0.9217 - recall_m: 0.8643 - val_loss: 0.3252 - val_acc: 0.9184 - val_precision_m: 0.9719 - val_recall_m: 0.8887\n",
      "Epoch 224/1000\n",
      " - 3s - loss: 0.3068 - acc: 0.9018 - precision_m: 0.9283 - recall_m: 0.8685 - val_loss: 0.3216 - val_acc: 0.9239 - val_precision_m: 0.9627 - val_recall_m: 0.8856\n",
      "Epoch 225/1000\n",
      " - 3s - loss: 0.3199 - acc: 0.8975 - precision_m: 0.9245 - recall_m: 0.8630 - val_loss: 0.3113 - val_acc: 0.9316 - val_precision_m: 0.9742 - val_recall_m: 0.8856\n",
      "Epoch 226/1000\n",
      " - 3s - loss: 0.3082 - acc: 0.8991 - precision_m: 0.9253 - recall_m: 0.8644 - val_loss: 0.3138 - val_acc: 0.9217 - val_precision_m: 0.9583 - val_recall_m: 0.8795\n",
      "Epoch 227/1000\n",
      " - 3s - loss: 0.3019 - acc: 0.9061 - precision_m: 0.9277 - recall_m: 0.8746 - val_loss: 0.3065 - val_acc: 0.9261 - val_precision_m: 0.9678 - val_recall_m: 0.8939\n",
      "Epoch 228/1000\n",
      " - 3s - loss: 0.3004 - acc: 0.9044 - precision_m: 0.9287 - recall_m: 0.8711 - val_loss: 0.3019 - val_acc: 0.9305 - val_precision_m: 0.9708 - val_recall_m: 0.8950\n",
      "Epoch 229/1000\n",
      " - 3s - loss: 0.3004 - acc: 0.9055 - precision_m: 0.9285 - recall_m: 0.8779 - val_loss: 0.3102 - val_acc: 0.9283 - val_precision_m: 0.9709 - val_recall_m: 0.8950\n",
      "Epoch 230/1000\n",
      " - 3s - loss: 0.2911 - acc: 0.9076 - precision_m: 0.9321 - recall_m: 0.8789 - val_loss: 0.2957 - val_acc: 0.9316 - val_precision_m: 0.9729 - val_recall_m: 0.8908\n",
      "Epoch 231/1000\n",
      " - 3s - loss: 0.2877 - acc: 0.9092 - precision_m: 0.9327 - recall_m: 0.8790 - val_loss: 0.2928 - val_acc: 0.9416 - val_precision_m: 0.9709 - val_recall_m: 0.9115\n",
      "Epoch 232/1000\n",
      " - 3s - loss: 0.2865 - acc: 0.9086 - precision_m: 0.9299 - recall_m: 0.8824 - val_loss: 0.2855 - val_acc: 0.9383 - val_precision_m: 0.9721 - val_recall_m: 0.9002\n",
      "Epoch 233/1000\n",
      " - 3s - loss: 0.2875 - acc: 0.9104 - precision_m: 0.9315 - recall_m: 0.8824 - val_loss: 0.2804 - val_acc: 0.9350 - val_precision_m: 0.9637 - val_recall_m: 0.9054\n",
      "Epoch 234/1000\n",
      " - 3s - loss: 0.2659 - acc: 0.9140 - precision_m: 0.9370 - recall_m: 0.8889 - val_loss: 0.2617 - val_acc: 0.9460 - val_precision_m: 0.9800 - val_recall_m: 0.9158\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 235/1000\n",
      " - 3s - loss: 0.2810 - acc: 0.9099 - precision_m: 0.9352 - recall_m: 0.8853 - val_loss: 0.2639 - val_acc: 0.9416 - val_precision_m: 0.9744 - val_recall_m: 0.9146\n",
      "Epoch 236/1000\n",
      " - 3s - loss: 0.2711 - acc: 0.9087 - precision_m: 0.9312 - recall_m: 0.8843 - val_loss: 0.2646 - val_acc: 0.9416 - val_precision_m: 0.9709 - val_recall_m: 0.9044\n",
      "Epoch 237/1000\n",
      " - 3s - loss: 0.2718 - acc: 0.9174 - precision_m: 0.9379 - recall_m: 0.8893 - val_loss: 0.2563 - val_acc: 0.9504 - val_precision_m: 0.9724 - val_recall_m: 0.9127\n",
      "Epoch 238/1000\n",
      " - 3s - loss: 0.2805 - acc: 0.9104 - precision_m: 0.9291 - recall_m: 0.8860 - val_loss: 0.2538 - val_acc: 0.9460 - val_precision_m: 0.9791 - val_recall_m: 0.9200\n",
      "Epoch 239/1000\n",
      " - 3s - loss: 0.2690 - acc: 0.9130 - precision_m: 0.9348 - recall_m: 0.8910 - val_loss: 0.2484 - val_acc: 0.9504 - val_precision_m: 0.9780 - val_recall_m: 0.9302\n",
      "Epoch 240/1000\n",
      " - 3s - loss: 0.2673 - acc: 0.9172 - precision_m: 0.9364 - recall_m: 0.8933 - val_loss: 0.2474 - val_acc: 0.9471 - val_precision_m: 0.9801 - val_recall_m: 0.9252\n",
      "Epoch 241/1000\n",
      " - 3s - loss: 0.2671 - acc: 0.9199 - precision_m: 0.9386 - recall_m: 0.8962 - val_loss: 0.2421 - val_acc: 0.9504 - val_precision_m: 0.9770 - val_recall_m: 0.9281\n",
      "Epoch 242/1000\n",
      " - 3s - loss: 0.2637 - acc: 0.9146 - precision_m: 0.9349 - recall_m: 0.8899 - val_loss: 0.2369 - val_acc: 0.9570 - val_precision_m: 0.9780 - val_recall_m: 0.9375\n",
      "Epoch 243/1000\n",
      " - 3s - loss: 0.2523 - acc: 0.9191 - precision_m: 0.9395 - recall_m: 0.8950 - val_loss: 0.2273 - val_acc: 0.9548 - val_precision_m: 0.9781 - val_recall_m: 0.9273\n",
      "Epoch 244/1000\n",
      " - 3s - loss: 0.2395 - acc: 0.9255 - precision_m: 0.9437 - recall_m: 0.9031 - val_loss: 0.2231 - val_acc: 0.9526 - val_precision_m: 0.9816 - val_recall_m: 0.9417\n",
      "Epoch 245/1000\n",
      " - 3s - loss: 0.2597 - acc: 0.9178 - precision_m: 0.9372 - recall_m: 0.8976 - val_loss: 0.2199 - val_acc: 0.9625 - val_precision_m: 0.9869 - val_recall_m: 0.9469\n",
      "Epoch 246/1000\n",
      " - 3s - loss: 0.2375 - acc: 0.9250 - precision_m: 0.9411 - recall_m: 0.9058 - val_loss: 0.2302 - val_acc: 0.9515 - val_precision_m: 0.9792 - val_recall_m: 0.9396\n",
      "Epoch 247/1000\n",
      " - 3s - loss: 0.2364 - acc: 0.9262 - precision_m: 0.9410 - recall_m: 0.9055 - val_loss: 0.2250 - val_acc: 0.9526 - val_precision_m: 0.9802 - val_recall_m: 0.9365\n",
      "Epoch 248/1000\n",
      " - 3s - loss: 0.2414 - acc: 0.9244 - precision_m: 0.9414 - recall_m: 0.9037 - val_loss: 0.2110 - val_acc: 0.9603 - val_precision_m: 0.9848 - val_recall_m: 0.9490\n",
      "Epoch 249/1000\n",
      " - 3s - loss: 0.2378 - acc: 0.9262 - precision_m: 0.9442 - recall_m: 0.9080 - val_loss: 0.2138 - val_acc: 0.9592 - val_precision_m: 0.9834 - val_recall_m: 0.9333\n",
      "Epoch 250/1000\n",
      " - 3s - loss: 0.2325 - acc: 0.9279 - precision_m: 0.9434 - recall_m: 0.9099 - val_loss: 0.2164 - val_acc: 0.9559 - val_precision_m: 0.9834 - val_recall_m: 0.9354\n",
      "Epoch 251/1000\n",
      " - 3s - loss: 0.2221 - acc: 0.9310 - precision_m: 0.9464 - recall_m: 0.9137 - val_loss: 0.2012 - val_acc: 0.9548 - val_precision_m: 0.9742 - val_recall_m: 0.9419\n",
      "Epoch 252/1000\n",
      " - 3s - loss: 0.2144 - acc: 0.9360 - precision_m: 0.9504 - recall_m: 0.9167 - val_loss: 0.1973 - val_acc: 0.9647 - val_precision_m: 0.9847 - val_recall_m: 0.9438\n",
      "Epoch 253/1000\n",
      " - 3s - loss: 0.2328 - acc: 0.9267 - precision_m: 0.9438 - recall_m: 0.9092 - val_loss: 0.1919 - val_acc: 0.9647 - val_precision_m: 0.9804 - val_recall_m: 0.9439\n",
      "Epoch 254/1000\n",
      " - 3s - loss: 0.2180 - acc: 0.9315 - precision_m: 0.9490 - recall_m: 0.9160 - val_loss: 0.1984 - val_acc: 0.9548 - val_precision_m: 0.9782 - val_recall_m: 0.9417\n",
      "Epoch 255/1000\n",
      " - 3s - loss: 0.2125 - acc: 0.9333 - precision_m: 0.9476 - recall_m: 0.9148 - val_loss: 0.1942 - val_acc: 0.9592 - val_precision_m: 0.9763 - val_recall_m: 0.9500\n",
      "Epoch 256/1000\n",
      " - 3s - loss: 0.2019 - acc: 0.9382 - precision_m: 0.9511 - recall_m: 0.9222 - val_loss: 0.1907 - val_acc: 0.9625 - val_precision_m: 0.9792 - val_recall_m: 0.9417\n",
      "Epoch 257/1000\n",
      " - 3s - loss: 0.2133 - acc: 0.9341 - precision_m: 0.9461 - recall_m: 0.9156 - val_loss: 0.1842 - val_acc: 0.9614 - val_precision_m: 0.9794 - val_recall_m: 0.9521\n",
      "Epoch 258/1000\n",
      " - 3s - loss: 0.2103 - acc: 0.9343 - precision_m: 0.9466 - recall_m: 0.9161 - val_loss: 0.1884 - val_acc: 0.9526 - val_precision_m: 0.9794 - val_recall_m: 0.9490\n",
      "Epoch 259/1000\n",
      " - 3s - loss: 0.2030 - acc: 0.9374 - precision_m: 0.9512 - recall_m: 0.9205 - val_loss: 0.1689 - val_acc: 0.9691 - val_precision_m: 0.9861 - val_recall_m: 0.9594\n",
      "Epoch 260/1000\n",
      " - 3s - loss: 0.2118 - acc: 0.9355 - precision_m: 0.9496 - recall_m: 0.9189 - val_loss: 0.1833 - val_acc: 0.9515 - val_precision_m: 0.9723 - val_recall_m: 0.9521\n",
      "Epoch 261/1000\n",
      " - 3s - loss: 0.1994 - acc: 0.9362 - precision_m: 0.9512 - recall_m: 0.9232 - val_loss: 0.1725 - val_acc: 0.9603 - val_precision_m: 0.9796 - val_recall_m: 0.9542\n",
      "Epoch 262/1000\n",
      " - 3s - loss: 0.2037 - acc: 0.9355 - precision_m: 0.9481 - recall_m: 0.9191 - val_loss: 0.1694 - val_acc: 0.9724 - val_precision_m: 0.9871 - val_recall_m: 0.9594\n",
      "Epoch 263/1000\n",
      " - 3s - loss: 0.1980 - acc: 0.9335 - precision_m: 0.9492 - recall_m: 0.9171 - val_loss: 0.1656 - val_acc: 0.9669 - val_precision_m: 0.9869 - val_recall_m: 0.9563\n",
      "Epoch 264/1000\n",
      " - 3s - loss: 0.2028 - acc: 0.9376 - precision_m: 0.9513 - recall_m: 0.9208 - val_loss: 0.1558 - val_acc: 0.9669 - val_precision_m: 0.9871 - val_recall_m: 0.9656\n",
      "Epoch 265/1000\n",
      " - 3s - loss: 0.2023 - acc: 0.9387 - precision_m: 0.9517 - recall_m: 0.9227 - val_loss: 0.1669 - val_acc: 0.9691 - val_precision_m: 0.9828 - val_recall_m: 0.9594\n",
      "Epoch 266/1000\n",
      " - 3s - loss: 0.1984 - acc: 0.9358 - precision_m: 0.9491 - recall_m: 0.9197 - val_loss: 0.1575 - val_acc: 0.9680 - val_precision_m: 0.9871 - val_recall_m: 0.9615\n",
      "Epoch 267/1000\n",
      " - 3s - loss: 0.1896 - acc: 0.9423 - precision_m: 0.9555 - recall_m: 0.9301 - val_loss: 0.1557 - val_acc: 0.9669 - val_precision_m: 0.9850 - val_recall_m: 0.9563\n",
      "Epoch 268/1000\n",
      " - 3s - loss: 0.1934 - acc: 0.9398 - precision_m: 0.9514 - recall_m: 0.9253 - val_loss: 0.1527 - val_acc: 0.9702 - val_precision_m: 0.9893 - val_recall_m: 0.9635\n",
      "Epoch 269/1000\n",
      " - 3s - loss: 0.1741 - acc: 0.9447 - precision_m: 0.9589 - recall_m: 0.9307 - val_loss: 0.1464 - val_acc: 0.9735 - val_precision_m: 0.9892 - val_recall_m: 0.9656\n",
      "Epoch 270/1000\n",
      " - 3s - loss: 0.1909 - acc: 0.9404 - precision_m: 0.9532 - recall_m: 0.9269 - val_loss: 0.1570 - val_acc: 0.9614 - val_precision_m: 0.9775 - val_recall_m: 0.9573\n",
      "Epoch 271/1000\n",
      " - 3s - loss: 0.1888 - acc: 0.9414 - precision_m: 0.9539 - recall_m: 0.9309 - val_loss: 0.1479 - val_acc: 0.9702 - val_precision_m: 0.9882 - val_recall_m: 0.9635\n",
      "Epoch 272/1000\n",
      " - 3s - loss: 0.1845 - acc: 0.9445 - precision_m: 0.9556 - recall_m: 0.9316 - val_loss: 0.1482 - val_acc: 0.9713 - val_precision_m: 0.9840 - val_recall_m: 0.9635\n",
      "Epoch 273/1000\n",
      " - 3s - loss: 0.1878 - acc: 0.9406 - precision_m: 0.9527 - recall_m: 0.9287 - val_loss: 0.1376 - val_acc: 0.9779 - val_precision_m: 0.9884 - val_recall_m: 0.9708\n",
      "Epoch 274/1000\n",
      " - 3s - loss: 0.1789 - acc: 0.9447 - precision_m: 0.9573 - recall_m: 0.9329 - val_loss: 0.1373 - val_acc: 0.9757 - val_precision_m: 0.9883 - val_recall_m: 0.9729\n",
      "Epoch 275/1000\n",
      " - 3s - loss: 0.1689 - acc: 0.9468 - precision_m: 0.9567 - recall_m: 0.9341 - val_loss: 0.1352 - val_acc: 0.9713 - val_precision_m: 0.9863 - val_recall_m: 0.9698\n",
      "Epoch 276/1000\n",
      " - 3s - loss: 0.1801 - acc: 0.9442 - precision_m: 0.9556 - recall_m: 0.9324 - val_loss: 0.1432 - val_acc: 0.9702 - val_precision_m: 0.9850 - val_recall_m: 0.9615\n",
      "Epoch 277/1000\n",
      " - 3s - loss: 0.1785 - acc: 0.9445 - precision_m: 0.9566 - recall_m: 0.9347 - val_loss: 0.1407 - val_acc: 0.9680 - val_precision_m: 0.9840 - val_recall_m: 0.9635\n",
      "Epoch 278/1000\n",
      " - 3s - loss: 0.1668 - acc: 0.9466 - precision_m: 0.9564 - recall_m: 0.9352 - val_loss: 0.1293 - val_acc: 0.9757 - val_precision_m: 0.9861 - val_recall_m: 0.9635\n",
      "Epoch 279/1000\n",
      " - 3s - loss: 0.1789 - acc: 0.9426 - precision_m: 0.9535 - recall_m: 0.9302 - val_loss: 0.1260 - val_acc: 0.9669 - val_precision_m: 0.9840 - val_recall_m: 0.9656\n",
      "Epoch 280/1000\n",
      " - 3s - loss: 0.1657 - acc: 0.9493 - precision_m: 0.9600 - recall_m: 0.9387 - val_loss: 0.1279 - val_acc: 0.9735 - val_precision_m: 0.9872 - val_recall_m: 0.9688\n",
      "Epoch 281/1000\n",
      " - 3s - loss: 0.1638 - acc: 0.9495 - precision_m: 0.9592 - recall_m: 0.9395 - val_loss: 0.1265 - val_acc: 0.9724 - val_precision_m: 0.9882 - val_recall_m: 0.9667\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 282/1000\n",
      " - 3s - loss: 0.1589 - acc: 0.9532 - precision_m: 0.9624 - recall_m: 0.9393 - val_loss: 0.1213 - val_acc: 0.9779 - val_precision_m: 0.9883 - val_recall_m: 0.9708\n",
      "Epoch 283/1000\n",
      " - 3s - loss: 0.1582 - acc: 0.9495 - precision_m: 0.9582 - recall_m: 0.9376 - val_loss: 0.1200 - val_acc: 0.9791 - val_precision_m: 0.9957 - val_recall_m: 0.9708\n",
      "Epoch 284/1000\n",
      " - 3s - loss: 0.1729 - acc: 0.9440 - precision_m: 0.9546 - recall_m: 0.9333 - val_loss: 0.1221 - val_acc: 0.9757 - val_precision_m: 0.9872 - val_recall_m: 0.9708\n",
      "Epoch 285/1000\n",
      " - 3s - loss: 0.1567 - acc: 0.9537 - precision_m: 0.9627 - recall_m: 0.9427 - val_loss: 0.1164 - val_acc: 0.9724 - val_precision_m: 0.9894 - val_recall_m: 0.9698\n",
      "Epoch 286/1000\n",
      " - 3s - loss: 0.1535 - acc: 0.9517 - precision_m: 0.9614 - recall_m: 0.9421 - val_loss: 0.1180 - val_acc: 0.9735 - val_precision_m: 0.9872 - val_recall_m: 0.9667\n",
      "Epoch 287/1000\n",
      " - 3s - loss: 0.1556 - acc: 0.9526 - precision_m: 0.9621 - recall_m: 0.9415 - val_loss: 0.1181 - val_acc: 0.9768 - val_precision_m: 0.9925 - val_recall_m: 0.9708\n",
      "Epoch 288/1000\n",
      " - 3s - loss: 0.1603 - acc: 0.9495 - precision_m: 0.9592 - recall_m: 0.9386 - val_loss: 0.1198 - val_acc: 0.9713 - val_precision_m: 0.9882 - val_recall_m: 0.9688\n",
      "Epoch 289/1000\n",
      " - 3s - loss: 0.1582 - acc: 0.9499 - precision_m: 0.9585 - recall_m: 0.9393 - val_loss: 0.1151 - val_acc: 0.9824 - val_precision_m: 0.9903 - val_recall_m: 0.9708\n",
      "Epoch 290/1000\n",
      " - 3s - loss: 0.1468 - acc: 0.9542 - precision_m: 0.9623 - recall_m: 0.9441 - val_loss: 0.1207 - val_acc: 0.9735 - val_precision_m: 0.9841 - val_recall_m: 0.9677\n",
      "Epoch 291/1000\n",
      " - 3s - loss: 0.1518 - acc: 0.9507 - precision_m: 0.9604 - recall_m: 0.9408 - val_loss: 0.1227 - val_acc: 0.9724 - val_precision_m: 0.9839 - val_recall_m: 0.9625\n",
      "Epoch 292/1000\n",
      " - 3s - loss: 0.1550 - acc: 0.9531 - precision_m: 0.9632 - recall_m: 0.9445 - val_loss: 0.1025 - val_acc: 0.9802 - val_precision_m: 0.9958 - val_recall_m: 0.9792\n",
      "Epoch 293/1000\n",
      " - 3s - loss: 0.1428 - acc: 0.9582 - precision_m: 0.9638 - recall_m: 0.9484 - val_loss: 0.1071 - val_acc: 0.9779 - val_precision_m: 0.9935 - val_recall_m: 0.9760\n",
      "Epoch 294/1000\n",
      " - 3s - loss: 0.1517 - acc: 0.9542 - precision_m: 0.9629 - recall_m: 0.9458 - val_loss: 0.0952 - val_acc: 0.9868 - val_precision_m: 0.9989 - val_recall_m: 0.9802\n",
      "Epoch 295/1000\n",
      " - 3s - loss: 0.1452 - acc: 0.9558 - precision_m: 0.9641 - recall_m: 0.9473 - val_loss: 0.1023 - val_acc: 0.9802 - val_precision_m: 0.9915 - val_recall_m: 0.9750\n",
      "Epoch 296/1000\n",
      " - 3s - loss: 0.1530 - acc: 0.9526 - precision_m: 0.9616 - recall_m: 0.9434 - val_loss: 0.1030 - val_acc: 0.9868 - val_precision_m: 0.9968 - val_recall_m: 0.9781\n",
      "Epoch 297/1000\n",
      " - 3s - loss: 0.1436 - acc: 0.9567 - precision_m: 0.9644 - recall_m: 0.9479 - val_loss: 0.1014 - val_acc: 0.9835 - val_precision_m: 0.9915 - val_recall_m: 0.9750\n",
      "Epoch 298/1000\n",
      " - 3s - loss: 0.1427 - acc: 0.9575 - precision_m: 0.9650 - recall_m: 0.9484 - val_loss: 0.0919 - val_acc: 0.9857 - val_precision_m: 0.9947 - val_recall_m: 0.9792\n",
      "Epoch 299/1000\n",
      " - 3s - loss: 0.1408 - acc: 0.9532 - precision_m: 0.9612 - recall_m: 0.9459 - val_loss: 0.0935 - val_acc: 0.9879 - val_precision_m: 0.9926 - val_recall_m: 0.9823\n",
      "Epoch 300/1000\n",
      " - 3s - loss: 0.1429 - acc: 0.9551 - precision_m: 0.9623 - recall_m: 0.9464 - val_loss: 0.0916 - val_acc: 0.9846 - val_precision_m: 0.9957 - val_recall_m: 0.9802\n",
      "Epoch 301/1000\n",
      " - 3s - loss: 0.1341 - acc: 0.9581 - precision_m: 0.9647 - recall_m: 0.9493 - val_loss: 0.0983 - val_acc: 0.9813 - val_precision_m: 0.9936 - val_recall_m: 0.9771\n",
      "Epoch 302/1000\n",
      " - 3s - loss: 0.1393 - acc: 0.9575 - precision_m: 0.9647 - recall_m: 0.9497 - val_loss: 0.0851 - val_acc: 0.9901 - val_precision_m: 0.9958 - val_recall_m: 0.9854\n",
      "Epoch 303/1000\n",
      " - 3s - loss: 0.1311 - acc: 0.9593 - precision_m: 0.9664 - recall_m: 0.9508 - val_loss: 0.0865 - val_acc: 0.9868 - val_precision_m: 0.9968 - val_recall_m: 0.9833\n",
      "Epoch 304/1000\n",
      " - 3s - loss: 0.1386 - acc: 0.9567 - precision_m: 0.9639 - recall_m: 0.9484 - val_loss: 0.0880 - val_acc: 0.9912 - val_precision_m: 0.9947 - val_recall_m: 0.9823\n",
      "Epoch 305/1000\n",
      " - 3s - loss: 0.1320 - acc: 0.9605 - precision_m: 0.9671 - recall_m: 0.9531 - val_loss: 0.0828 - val_acc: 0.9901 - val_precision_m: 0.9947 - val_recall_m: 0.9844\n",
      "Epoch 306/1000\n",
      " - 3s - loss: 0.1345 - acc: 0.9591 - precision_m: 0.9654 - recall_m: 0.9518 - val_loss: 0.0895 - val_acc: 0.9857 - val_precision_m: 0.9916 - val_recall_m: 0.9823\n",
      "Epoch 307/1000\n",
      " - 3s - loss: 0.1317 - acc: 0.9587 - precision_m: 0.9649 - recall_m: 0.9514 - val_loss: 0.0816 - val_acc: 0.9890 - val_precision_m: 0.9979 - val_recall_m: 0.9833\n",
      "Epoch 308/1000\n",
      " - 3s - loss: 0.1371 - acc: 0.9575 - precision_m: 0.9636 - recall_m: 0.9500 - val_loss: 0.0833 - val_acc: 0.9879 - val_precision_m: 0.9884 - val_recall_m: 0.9823\n",
      "Epoch 309/1000\n",
      " - 3s - loss: 0.1296 - acc: 0.9607 - precision_m: 0.9668 - recall_m: 0.9519 - val_loss: 0.0831 - val_acc: 0.9934 - val_precision_m: 0.9968 - val_recall_m: 0.9844\n",
      "Epoch 310/1000\n",
      " - 3s - loss: 0.1304 - acc: 0.9609 - precision_m: 0.9679 - recall_m: 0.9536 - val_loss: 0.0840 - val_acc: 0.9923 - val_precision_m: 0.9979 - val_recall_m: 0.9833\n",
      "Epoch 311/1000\n",
      " - 3s - loss: 0.1289 - acc: 0.9616 - precision_m: 0.9676 - recall_m: 0.9548 - val_loss: 0.0816 - val_acc: 0.9912 - val_precision_m: 0.9989 - val_recall_m: 0.9854\n",
      "Epoch 312/1000\n",
      " - 3s - loss: 0.1158 - acc: 0.9650 - precision_m: 0.9709 - recall_m: 0.9591 - val_loss: 0.0807 - val_acc: 0.9912 - val_precision_m: 0.9937 - val_recall_m: 0.9865\n",
      "Epoch 313/1000\n",
      " - 3s - loss: 0.1250 - acc: 0.9626 - precision_m: 0.9690 - recall_m: 0.9546 - val_loss: 0.0803 - val_acc: 0.9923 - val_precision_m: 0.9979 - val_recall_m: 0.9844\n",
      "Epoch 314/1000\n",
      " - 3s - loss: 0.1163 - acc: 0.9615 - precision_m: 0.9676 - recall_m: 0.9563 - val_loss: 0.0829 - val_acc: 0.9912 - val_precision_m: 0.9968 - val_recall_m: 0.9833\n",
      "Epoch 315/1000\n",
      " - 3s - loss: 0.1275 - acc: 0.9583 - precision_m: 0.9669 - recall_m: 0.9518 - val_loss: 0.0856 - val_acc: 0.9945 - val_precision_m: 1.0000 - val_recall_m: 0.9833\n",
      "Epoch 316/1000\n",
      " - 3s - loss: 0.1116 - acc: 0.9654 - precision_m: 0.9716 - recall_m: 0.9585 - val_loss: 0.0786 - val_acc: 0.9901 - val_precision_m: 0.9927 - val_recall_m: 0.9854\n",
      "Epoch 317/1000\n",
      " - 3s - loss: 0.1210 - acc: 0.9629 - precision_m: 0.9686 - recall_m: 0.9563 - val_loss: 0.0828 - val_acc: 0.9923 - val_precision_m: 0.9968 - val_recall_m: 0.9854\n",
      "Epoch 318/1000\n",
      " - 3s - loss: 0.1202 - acc: 0.9616 - precision_m: 0.9676 - recall_m: 0.9561 - val_loss: 0.0754 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9885\n",
      "Epoch 319/1000\n",
      " - 3s - loss: 0.1173 - acc: 0.9651 - precision_m: 0.9710 - recall_m: 0.9600 - val_loss: 0.0784 - val_acc: 0.9879 - val_precision_m: 0.9958 - val_recall_m: 0.9865\n",
      "Epoch 320/1000\n",
      " - 3s - loss: 0.1088 - acc: 0.9686 - precision_m: 0.9739 - recall_m: 0.9624 - val_loss: 0.0764 - val_acc: 0.9912 - val_precision_m: 0.9937 - val_recall_m: 0.9885\n",
      "Epoch 321/1000\n",
      " - 3s - loss: 0.1104 - acc: 0.9673 - precision_m: 0.9735 - recall_m: 0.9604 - val_loss: 0.0710 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9896\n",
      "Epoch 322/1000\n",
      " - 3s - loss: 0.1147 - acc: 0.9657 - precision_m: 0.9708 - recall_m: 0.9597 - val_loss: 0.0761 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9885\n",
      "Epoch 323/1000\n",
      " - 3s - loss: 0.1077 - acc: 0.9675 - precision_m: 0.9734 - recall_m: 0.9608 - val_loss: 0.0676 - val_acc: 0.9890 - val_precision_m: 0.9968 - val_recall_m: 0.9875\n",
      "Epoch 324/1000\n",
      " - 3s - loss: 0.1065 - acc: 0.9667 - precision_m: 0.9714 - recall_m: 0.9598 - val_loss: 0.0669 - val_acc: 0.9934 - val_precision_m: 0.9947 - val_recall_m: 0.9896\n",
      "Epoch 325/1000\n",
      " - 3s - loss: 0.1095 - acc: 0.9664 - precision_m: 0.9722 - recall_m: 0.9597 - val_loss: 0.0736 - val_acc: 0.9912 - val_precision_m: 0.9968 - val_recall_m: 0.9854\n",
      "Epoch 326/1000\n",
      " - 3s - loss: 0.1109 - acc: 0.9658 - precision_m: 0.9714 - recall_m: 0.9603 - val_loss: 0.0736 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9875\n",
      "Epoch 327/1000\n",
      " - 3s - loss: 0.1048 - acc: 0.9647 - precision_m: 0.9685 - recall_m: 0.9603 - val_loss: 0.0697 - val_acc: 0.9890 - val_precision_m: 0.9968 - val_recall_m: 0.9844\n",
      "Epoch 328/1000\n",
      " - 3s - loss: 0.1032 - acc: 0.9689 - precision_m: 0.9730 - recall_m: 0.9635 - val_loss: 0.0769 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9865\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 329/1000\n",
      " - 3s - loss: 0.1052 - acc: 0.9658 - precision_m: 0.9704 - recall_m: 0.9604 - val_loss: 0.0755 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 330/1000\n",
      " - 3s - loss: 0.1024 - acc: 0.9708 - precision_m: 0.9747 - recall_m: 0.9653 - val_loss: 0.0698 - val_acc: 0.9912 - val_precision_m: 0.9957 - val_recall_m: 0.9865\n",
      "Epoch 331/1000\n",
      " - 3s - loss: 0.1062 - acc: 0.9665 - precision_m: 0.9730 - recall_m: 0.9614 - val_loss: 0.0643 - val_acc: 0.9956 - val_precision_m: 0.9958 - val_recall_m: 0.9865\n",
      "Epoch 332/1000\n",
      " - 3s - loss: 0.1078 - acc: 0.9653 - precision_m: 0.9697 - recall_m: 0.9607 - val_loss: 0.0799 - val_acc: 0.9879 - val_precision_m: 0.9905 - val_recall_m: 0.9823\n",
      "Epoch 333/1000\n",
      " - 3s - loss: 0.1048 - acc: 0.9675 - precision_m: 0.9729 - recall_m: 0.9630 - val_loss: 0.0753 - val_acc: 0.9912 - val_precision_m: 0.9916 - val_recall_m: 0.9865\n",
      "Epoch 334/1000\n",
      " - 3s - loss: 0.1176 - acc: 0.9640 - precision_m: 0.9692 - recall_m: 0.9581 - val_loss: 0.0725 - val_acc: 0.9879 - val_precision_m: 0.9937 - val_recall_m: 0.9865\n",
      "Epoch 335/1000\n",
      " - 3s - loss: 0.1013 - acc: 0.9675 - precision_m: 0.9717 - recall_m: 0.9635 - val_loss: 0.0696 - val_acc: 0.9956 - val_precision_m: 0.9969 - val_recall_m: 0.9875\n",
      "Epoch 336/1000\n",
      " - 3s - loss: 0.0964 - acc: 0.9723 - precision_m: 0.9773 - recall_m: 0.9677 - val_loss: 0.0686 - val_acc: 0.9945 - val_precision_m: 0.9958 - val_recall_m: 0.9906\n",
      "Epoch 337/1000\n",
      " - 3s - loss: 0.0957 - acc: 0.9697 - precision_m: 0.9735 - recall_m: 0.9646 - val_loss: 0.0681 - val_acc: 0.9923 - val_precision_m: 0.9958 - val_recall_m: 0.9875\n",
      "Epoch 338/1000\n",
      " - 3s - loss: 0.1007 - acc: 0.9690 - precision_m: 0.9727 - recall_m: 0.9628 - val_loss: 0.0615 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9896\n",
      "Epoch 339/1000\n",
      " - 3s - loss: 0.1018 - acc: 0.9705 - precision_m: 0.9751 - recall_m: 0.9666 - val_loss: 0.0603 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9885\n",
      "Epoch 340/1000\n",
      " - 3s - loss: 0.0963 - acc: 0.9695 - precision_m: 0.9737 - recall_m: 0.9652 - val_loss: 0.0651 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9885\n",
      "Epoch 341/1000\n",
      " - 3s - loss: 0.1005 - acc: 0.9702 - precision_m: 0.9740 - recall_m: 0.9650 - val_loss: 0.0636 - val_acc: 0.9956 - val_precision_m: 1.0000 - val_recall_m: 0.9896\n",
      "Epoch 342/1000\n",
      " - 3s - loss: 0.0993 - acc: 0.9678 - precision_m: 0.9729 - recall_m: 0.9640 - val_loss: 0.0631 - val_acc: 0.9901 - val_precision_m: 0.9968 - val_recall_m: 0.9906\n",
      "Epoch 343/1000\n",
      " - 3s - loss: 0.0876 - acc: 0.9741 - precision_m: 0.9787 - recall_m: 0.9684 - val_loss: 0.0622 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 344/1000\n",
      " - 3s - loss: 0.0950 - acc: 0.9721 - precision_m: 0.9760 - recall_m: 0.9678 - val_loss: 0.0665 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 345/1000\n",
      " - 3s - loss: 0.0908 - acc: 0.9719 - precision_m: 0.9751 - recall_m: 0.9678 - val_loss: 0.0623 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9896\n",
      "Epoch 346/1000\n",
      " - 3s - loss: 0.0953 - acc: 0.9689 - precision_m: 0.9742 - recall_m: 0.9639 - val_loss: 0.0636 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 347/1000\n",
      " - 3s - loss: 0.0963 - acc: 0.9703 - precision_m: 0.9751 - recall_m: 0.9659 - val_loss: 0.0567 - val_acc: 0.9934 - val_precision_m: 1.0000 - val_recall_m: 0.9937\n",
      "Epoch 348/1000\n",
      " - 3s - loss: 0.0903 - acc: 0.9723 - precision_m: 0.9763 - recall_m: 0.9663 - val_loss: 0.0584 - val_acc: 0.9956 - val_precision_m: 0.9958 - val_recall_m: 0.9917\n",
      "Epoch 349/1000\n",
      " - 3s - loss: 0.0881 - acc: 0.9718 - precision_m: 0.9764 - recall_m: 0.9678 - val_loss: 0.0611 - val_acc: 0.9912 - val_precision_m: 0.9927 - val_recall_m: 0.9896\n",
      "Epoch 350/1000\n",
      " - 3s - loss: 0.0982 - acc: 0.9692 - precision_m: 0.9735 - recall_m: 0.9653 - val_loss: 0.0625 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9927\n",
      "Epoch 351/1000\n",
      " - 3s - loss: 0.0868 - acc: 0.9745 - precision_m: 0.9775 - recall_m: 0.9697 - val_loss: 0.0579 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 352/1000\n",
      " - 3s - loss: 0.0903 - acc: 0.9745 - precision_m: 0.9782 - recall_m: 0.9703 - val_loss: 0.0564 - val_acc: 0.9967 - val_precision_m: 0.9989 - val_recall_m: 0.9937\n",
      "Epoch 353/1000\n",
      " - 3s - loss: 0.0921 - acc: 0.9728 - precision_m: 0.9752 - recall_m: 0.9685 - val_loss: 0.0641 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9896\n",
      "Epoch 354/1000\n",
      " - 3s - loss: 0.0904 - acc: 0.9746 - precision_m: 0.9769 - recall_m: 0.9707 - val_loss: 0.0573 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 355/1000\n",
      " - 3s - loss: 0.0836 - acc: 0.9745 - precision_m: 0.9768 - recall_m: 0.9706 - val_loss: 0.0657 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9896\n",
      "Epoch 356/1000\n",
      " - 3s - loss: 0.1018 - acc: 0.9669 - precision_m: 0.9711 - recall_m: 0.9633 - val_loss: 0.0618 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 357/1000\n",
      " - 3s - loss: 0.0798 - acc: 0.9744 - precision_m: 0.9786 - recall_m: 0.9712 - val_loss: 0.0643 - val_acc: 0.9879 - val_precision_m: 0.9906 - val_recall_m: 0.9885\n",
      "Epoch 358/1000\n",
      " - 3s - loss: 0.0907 - acc: 0.9724 - precision_m: 0.9751 - recall_m: 0.9669 - val_loss: 0.0611 - val_acc: 0.9945 - val_precision_m: 0.9979 - val_recall_m: 0.9937\n",
      "Epoch 359/1000\n",
      " - 3s - loss: 0.0881 - acc: 0.9739 - precision_m: 0.9772 - recall_m: 0.9696 - val_loss: 0.0640 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9896\n",
      "Epoch 360/1000\n",
      " - 3s - loss: 0.0899 - acc: 0.9714 - precision_m: 0.9767 - recall_m: 0.9670 - val_loss: 0.0600 - val_acc: 0.9934 - val_precision_m: 0.9968 - val_recall_m: 0.9906\n",
      "Epoch 361/1000\n",
      " - 3s - loss: 0.0823 - acc: 0.9740 - precision_m: 0.9779 - recall_m: 0.9702 - val_loss: 0.0604 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 362/1000\n",
      " - 3s - loss: 0.0833 - acc: 0.9743 - precision_m: 0.9771 - recall_m: 0.9696 - val_loss: 0.0685 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 363/1000\n",
      " - 3s - loss: 0.0812 - acc: 0.9749 - precision_m: 0.9774 - recall_m: 0.9722 - val_loss: 0.0542 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 364/1000\n",
      " - 3s - loss: 0.0801 - acc: 0.9759 - precision_m: 0.9790 - recall_m: 0.9717 - val_loss: 0.0555 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9896\n",
      "Epoch 365/1000\n",
      " - 3s - loss: 0.0860 - acc: 0.9723 - precision_m: 0.9753 - recall_m: 0.9691 - val_loss: 0.0556 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 366/1000\n",
      " - 3s - loss: 0.0817 - acc: 0.9750 - precision_m: 0.9780 - recall_m: 0.9713 - val_loss: 0.0555 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 367/1000\n",
      " - 3s - loss: 0.0806 - acc: 0.9761 - precision_m: 0.9793 - recall_m: 0.9723 - val_loss: 0.0479 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 368/1000\n",
      " - 3s - loss: 0.0832 - acc: 0.9744 - precision_m: 0.9773 - recall_m: 0.9717 - val_loss: 0.0522 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9958\n",
      "Epoch 369/1000\n",
      " - 3s - loss: 0.0918 - acc: 0.9718 - precision_m: 0.9754 - recall_m: 0.9684 - val_loss: 0.0618 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 370/1000\n",
      " - 3s - loss: 0.0829 - acc: 0.9737 - precision_m: 0.9774 - recall_m: 0.9718 - val_loss: 0.0520 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 371/1000\n",
      " - 3s - loss: 0.0778 - acc: 0.9760 - precision_m: 0.9788 - recall_m: 0.9731 - val_loss: 0.0551 - val_acc: 0.9956 - val_precision_m: 0.9969 - val_recall_m: 0.9948\n",
      "Epoch 372/1000\n",
      " - 3s - loss: 0.0751 - acc: 0.9788 - precision_m: 0.9813 - recall_m: 0.9761 - val_loss: 0.0546 - val_acc: 0.9956 - val_precision_m: 0.9969 - val_recall_m: 0.9948\n",
      "Epoch 373/1000\n",
      " - 3s - loss: 0.0727 - acc: 0.9772 - precision_m: 0.9797 - recall_m: 0.9739 - val_loss: 0.0537 - val_acc: 0.9945 - val_precision_m: 0.9958 - val_recall_m: 0.9917\n",
      "Epoch 374/1000\n",
      " - 3s - loss: 0.0868 - acc: 0.9744 - precision_m: 0.9769 - recall_m: 0.9709 - val_loss: 0.0580 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 375/1000\n",
      " - 3s - loss: 0.0764 - acc: 0.9772 - precision_m: 0.9793 - recall_m: 0.9739 - val_loss: 0.0524 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 376/1000\n",
      " - 3s - loss: 0.0729 - acc: 0.9788 - precision_m: 0.9824 - recall_m: 0.9766 - val_loss: 0.0504 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 377/1000\n",
      " - 3s - loss: 0.0723 - acc: 0.9794 - precision_m: 0.9827 - recall_m: 0.9752 - val_loss: 0.0454 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 378/1000\n",
      " - 3s - loss: 0.0727 - acc: 0.9781 - precision_m: 0.9814 - recall_m: 0.9741 - val_loss: 0.0563 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 379/1000\n",
      " - 3s - loss: 0.0776 - acc: 0.9767 - precision_m: 0.9793 - recall_m: 0.9736 - val_loss: 0.0506 - val_acc: 0.9923 - val_precision_m: 0.9927 - val_recall_m: 0.9927\n",
      "Epoch 380/1000\n",
      " - 3s - loss: 0.0747 - acc: 0.9768 - precision_m: 0.9802 - recall_m: 0.9738 - val_loss: 0.0505 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 381/1000\n",
      " - 3s - loss: 0.0756 - acc: 0.9773 - precision_m: 0.9799 - recall_m: 0.9741 - val_loss: 0.0471 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 382/1000\n",
      " - 3s - loss: 0.0770 - acc: 0.9762 - precision_m: 0.9788 - recall_m: 0.9742 - val_loss: 0.0503 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9927\n",
      "Epoch 383/1000\n",
      " - 3s - loss: 0.0685 - acc: 0.9778 - precision_m: 0.9814 - recall_m: 0.9760 - val_loss: 0.0525 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 384/1000\n",
      " - 3s - loss: 0.0773 - acc: 0.9756 - precision_m: 0.9782 - recall_m: 0.9724 - val_loss: 0.0491 - val_acc: 0.9945 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 385/1000\n",
      " - 3s - loss: 0.0735 - acc: 0.9794 - precision_m: 0.9825 - recall_m: 0.9756 - val_loss: 0.0512 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 386/1000\n",
      " - 3s - loss: 0.0718 - acc: 0.9775 - precision_m: 0.9798 - recall_m: 0.9746 - val_loss: 0.0487 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 387/1000\n",
      " - 3s - loss: 0.0732 - acc: 0.9772 - precision_m: 0.9789 - recall_m: 0.9751 - val_loss: 0.0513 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 388/1000\n",
      " - 3s - loss: 0.0773 - acc: 0.9755 - precision_m: 0.9784 - recall_m: 0.9729 - val_loss: 0.0459 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 389/1000\n",
      " - 3s - loss: 0.0728 - acc: 0.9771 - precision_m: 0.9800 - recall_m: 0.9753 - val_loss: 0.0520 - val_acc: 0.9967 - val_precision_m: 0.9979 - val_recall_m: 0.9948\n",
      "Epoch 390/1000\n",
      " - 3s - loss: 0.0774 - acc: 0.9768 - precision_m: 0.9796 - recall_m: 0.9734 - val_loss: 0.0450 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 391/1000\n",
      " - 3s - loss: 0.0673 - acc: 0.9827 - precision_m: 0.9849 - recall_m: 0.9794 - val_loss: 0.0453 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 392/1000\n",
      " - 3s - loss: 0.0673 - acc: 0.9793 - precision_m: 0.9822 - recall_m: 0.9768 - val_loss: 0.0515 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 393/1000\n",
      " - 3s - loss: 0.0670 - acc: 0.9792 - precision_m: 0.9832 - recall_m: 0.9769 - val_loss: 0.0449 - val_acc: 0.9967 - val_precision_m: 1.0000 - val_recall_m: 0.9969\n",
      "Epoch 394/1000\n",
      " - 3s - loss: 0.0697 - acc: 0.9782 - precision_m: 0.9810 - recall_m: 0.9756 - val_loss: 0.0548 - val_acc: 0.9901 - val_precision_m: 0.9917 - val_recall_m: 0.9906\n",
      "Epoch 395/1000\n",
      " - 3s - loss: 0.0659 - acc: 0.9786 - precision_m: 0.9809 - recall_m: 0.9745 - val_loss: 0.0483 - val_acc: 0.9967 - val_precision_m: 1.0000 - val_recall_m: 0.9937\n",
      "Epoch 396/1000\n",
      " - 3s - loss: 0.0547 - acc: 0.9850 - precision_m: 0.9875 - recall_m: 0.9817 - val_loss: 0.0493 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 397/1000\n",
      " - 3s - loss: 0.0681 - acc: 0.9797 - precision_m: 0.9823 - recall_m: 0.9772 - val_loss: 0.0512 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 398/1000\n",
      " - 3s - loss: 0.0665 - acc: 0.9797 - precision_m: 0.9818 - recall_m: 0.9777 - val_loss: 0.0501 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 399/1000\n",
      " - 3s - loss: 0.0685 - acc: 0.9798 - precision_m: 0.9820 - recall_m: 0.9769 - val_loss: 0.0481 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 400/1000\n",
      " - 3s - loss: 0.0697 - acc: 0.9781 - precision_m: 0.9810 - recall_m: 0.9753 - val_loss: 0.0475 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 401/1000\n",
      " - 3s - loss: 0.0630 - acc: 0.9819 - precision_m: 0.9840 - recall_m: 0.9786 - val_loss: 0.0466 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 402/1000\n",
      " - 3s - loss: 0.0654 - acc: 0.9812 - precision_m: 0.9836 - recall_m: 0.9790 - val_loss: 0.0520 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 403/1000\n",
      " - 3s - loss: 0.0611 - acc: 0.9815 - precision_m: 0.9836 - recall_m: 0.9801 - val_loss: 0.0528 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 404/1000\n",
      " - 3s - loss: 0.0695 - acc: 0.9801 - precision_m: 0.9832 - recall_m: 0.9773 - val_loss: 0.0475 - val_acc: 0.9934 - val_precision_m: 0.9979 - val_recall_m: 0.9937\n",
      "Epoch 405/1000\n",
      " - 3s - loss: 0.0616 - acc: 0.9814 - precision_m: 0.9838 - recall_m: 0.9792 - val_loss: 0.0470 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 406/1000\n",
      " - 3s - loss: 0.0713 - acc: 0.9779 - precision_m: 0.9797 - recall_m: 0.9760 - val_loss: 0.0431 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9948\n",
      "Epoch 407/1000\n",
      " - 3s - loss: 0.0695 - acc: 0.9809 - precision_m: 0.9828 - recall_m: 0.9792 - val_loss: 0.0497 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 408/1000\n",
      " - 3s - loss: 0.0633 - acc: 0.9817 - precision_m: 0.9847 - recall_m: 0.9789 - val_loss: 0.0518 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 409/1000\n",
      " - 3s - loss: 0.0646 - acc: 0.9795 - precision_m: 0.9821 - recall_m: 0.9775 - val_loss: 0.0482 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 410/1000\n",
      " - 3s - loss: 0.0536 - acc: 0.9846 - precision_m: 0.9865 - recall_m: 0.9822 - val_loss: 0.0467 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 411/1000\n",
      " - 3s - loss: 0.0597 - acc: 0.9817 - precision_m: 0.9839 - recall_m: 0.9791 - val_loss: 0.0457 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 412/1000\n",
      " - 3s - loss: 0.0640 - acc: 0.9801 - precision_m: 0.9815 - recall_m: 0.9774 - val_loss: 0.0466 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 413/1000\n",
      " - 3s - loss: 0.0600 - acc: 0.9830 - precision_m: 0.9848 - recall_m: 0.9802 - val_loss: 0.0431 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9958\n",
      "Epoch 414/1000\n",
      " - 3s - loss: 0.0613 - acc: 0.9804 - precision_m: 0.9822 - recall_m: 0.9792 - val_loss: 0.0440 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 415/1000\n",
      " - 3s - loss: 0.0665 - acc: 0.9771 - precision_m: 0.9788 - recall_m: 0.9746 - val_loss: 0.0405 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 416/1000\n",
      " - 3s - loss: 0.0540 - acc: 0.9841 - precision_m: 0.9866 - recall_m: 0.9822 - val_loss: 0.0431 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 417/1000\n",
      " - 3s - loss: 0.0596 - acc: 0.9801 - precision_m: 0.9839 - recall_m: 0.9775 - val_loss: 0.0496 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 418/1000\n",
      " - 3s - loss: 0.0659 - acc: 0.9795 - precision_m: 0.9814 - recall_m: 0.9772 - val_loss: 0.0533 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 419/1000\n",
      " - 3s - loss: 0.0607 - acc: 0.9824 - precision_m: 0.9838 - recall_m: 0.9800 - val_loss: 0.0515 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 420/1000\n",
      " - 3s - loss: 0.0622 - acc: 0.9821 - precision_m: 0.9833 - recall_m: 0.9789 - val_loss: 0.0504 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 421/1000\n",
      " - 3s - loss: 0.0559 - acc: 0.9838 - precision_m: 0.9858 - recall_m: 0.9806 - val_loss: 0.0497 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 422/1000\n",
      " - 3s - loss: 0.0632 - acc: 0.9799 - precision_m: 0.9816 - recall_m: 0.9783 - val_loss: 0.0528 - val_acc: 0.9912 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 423/1000\n",
      " - 3s - loss: 0.0590 - acc: 0.9812 - precision_m: 0.9838 - recall_m: 0.9799 - val_loss: 0.0466 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 424/1000\n",
      " - 3s - loss: 0.0633 - acc: 0.9810 - precision_m: 0.9832 - recall_m: 0.9788 - val_loss: 0.0432 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 425/1000\n",
      " - 3s - loss: 0.0532 - acc: 0.9838 - precision_m: 0.9856 - recall_m: 0.9823 - val_loss: 0.0379 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 426/1000\n",
      " - 3s - loss: 0.0536 - acc: 0.9827 - precision_m: 0.9848 - recall_m: 0.9814 - val_loss: 0.0534 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 427/1000\n",
      " - 3s - loss: 0.0613 - acc: 0.9805 - precision_m: 0.9827 - recall_m: 0.9786 - val_loss: 0.0442 - val_acc: 0.9967 - val_precision_m: 0.9989 - val_recall_m: 0.9969\n",
      "Epoch 428/1000\n",
      " - 3s - loss: 0.0576 - acc: 0.9822 - precision_m: 0.9833 - recall_m: 0.9800 - val_loss: 0.0500 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 429/1000\n",
      " - 3s - loss: 0.0561 - acc: 0.9828 - precision_m: 0.9849 - recall_m: 0.9811 - val_loss: 0.0483 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 430/1000\n",
      " - 3s - loss: 0.0619 - acc: 0.9817 - precision_m: 0.9844 - recall_m: 0.9792 - val_loss: 0.0470 - val_acc: 0.9967 - val_precision_m: 0.9989 - val_recall_m: 0.9969\n",
      "Epoch 431/1000\n",
      " - 3s - loss: 0.0620 - acc: 0.9806 - precision_m: 0.9824 - recall_m: 0.9778 - val_loss: 0.0501 - val_acc: 0.9901 - val_precision_m: 0.9917 - val_recall_m: 0.9906\n",
      "Epoch 432/1000\n",
      " - 3s - loss: 0.0541 - acc: 0.9853 - precision_m: 0.9865 - recall_m: 0.9839 - val_loss: 0.0500 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 433/1000\n",
      " - 3s - loss: 0.0606 - acc: 0.9836 - precision_m: 0.9849 - recall_m: 0.9813 - val_loss: 0.0496 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 434/1000\n",
      " - 3s - loss: 0.0571 - acc: 0.9811 - precision_m: 0.9842 - recall_m: 0.9791 - val_loss: 0.0450 - val_acc: 0.9945 - val_precision_m: 0.9979 - val_recall_m: 0.9948\n",
      "Epoch 435/1000\n",
      " - 3s - loss: 0.0560 - acc: 0.9824 - precision_m: 0.9846 - recall_m: 0.9810 - val_loss: 0.0468 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 436/1000\n",
      " - 3s - loss: 0.0546 - acc: 0.9838 - precision_m: 0.9856 - recall_m: 0.9811 - val_loss: 0.0460 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 437/1000\n",
      " - 3s - loss: 0.0511 - acc: 0.9854 - precision_m: 0.9867 - recall_m: 0.9836 - val_loss: 0.0408 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 438/1000\n",
      " - 3s - loss: 0.0632 - acc: 0.9788 - precision_m: 0.9808 - recall_m: 0.9772 - val_loss: 0.0428 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 439/1000\n",
      " - 3s - loss: 0.0548 - acc: 0.9846 - precision_m: 0.9851 - recall_m: 0.9832 - val_loss: 0.0412 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 440/1000\n",
      " - 3s - loss: 0.0601 - acc: 0.9815 - precision_m: 0.9834 - recall_m: 0.9794 - val_loss: 0.0438 - val_acc: 0.9945 - val_precision_m: 0.9948 - val_recall_m: 0.9917\n",
      "Epoch 441/1000\n",
      " - 3s - loss: 0.0569 - acc: 0.9814 - precision_m: 0.9836 - recall_m: 0.9794 - val_loss: 0.0491 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 442/1000\n",
      " - 3s - loss: 0.0492 - acc: 0.9850 - precision_m: 0.9868 - recall_m: 0.9827 - val_loss: 0.0558 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9917\n",
      "Epoch 443/1000\n",
      " - 3s - loss: 0.0526 - acc: 0.9841 - precision_m: 0.9855 - recall_m: 0.9821 - val_loss: 0.0502 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 444/1000\n",
      " - 3s - loss: 0.0479 - acc: 0.9875 - precision_m: 0.9891 - recall_m: 0.9852 - val_loss: 0.0510 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 445/1000\n",
      " - 3s - loss: 0.0519 - acc: 0.9839 - precision_m: 0.9860 - recall_m: 0.9827 - val_loss: 0.0571 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 446/1000\n",
      " - 3s - loss: 0.0479 - acc: 0.9835 - precision_m: 0.9843 - recall_m: 0.9821 - val_loss: 0.0593 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 447/1000\n",
      " - 3s - loss: 0.0587 - acc: 0.9826 - precision_m: 0.9835 - recall_m: 0.9805 - val_loss: 0.0504 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9917\n",
      "Epoch 448/1000\n",
      " - 3s - loss: 0.0484 - acc: 0.9874 - precision_m: 0.9887 - recall_m: 0.9863 - val_loss: 0.0478 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 449/1000\n",
      " - 3s - loss: 0.0501 - acc: 0.9843 - precision_m: 0.9858 - recall_m: 0.9830 - val_loss: 0.0524 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 450/1000\n",
      " - 3s - loss: 0.0466 - acc: 0.9852 - precision_m: 0.9870 - recall_m: 0.9840 - val_loss: 0.0445 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 451/1000\n",
      " - 3s - loss: 0.0467 - acc: 0.9870 - precision_m: 0.9885 - recall_m: 0.9857 - val_loss: 0.0457 - val_acc: 0.9901 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 452/1000\n",
      " - 3s - loss: 0.0482 - acc: 0.9854 - precision_m: 0.9862 - recall_m: 0.9839 - val_loss: 0.0468 - val_acc: 0.9923 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 453/1000\n",
      " - 3s - loss: 0.0509 - acc: 0.9848 - precision_m: 0.9863 - recall_m: 0.9830 - val_loss: 0.0487 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 454/1000\n",
      " - 3s - loss: 0.0478 - acc: 0.9852 - precision_m: 0.9865 - recall_m: 0.9839 - val_loss: 0.0458 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 455/1000\n",
      " - 3s - loss: 0.0475 - acc: 0.9846 - precision_m: 0.9856 - recall_m: 0.9838 - val_loss: 0.0490 - val_acc: 0.9967 - val_precision_m: 1.0000 - val_recall_m: 0.9937\n",
      "Epoch 456/1000\n",
      " - 3s - loss: 0.0492 - acc: 0.9848 - precision_m: 0.9859 - recall_m: 0.9829 - val_loss: 0.0422 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 457/1000\n",
      " - 3s - loss: 0.0491 - acc: 0.9857 - precision_m: 0.9865 - recall_m: 0.9840 - val_loss: 0.0456 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 458/1000\n",
      " - 3s - loss: 0.0588 - acc: 0.9810 - precision_m: 0.9830 - recall_m: 0.9796 - val_loss: 0.0416 - val_acc: 0.9934 - val_precision_m: 0.9958 - val_recall_m: 0.9937\n",
      "Epoch 459/1000\n",
      " - 3s - loss: 0.0467 - acc: 0.9858 - precision_m: 0.9868 - recall_m: 0.9854 - val_loss: 0.0460 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 460/1000\n",
      " - 3s - loss: 0.0478 - acc: 0.9848 - precision_m: 0.9866 - recall_m: 0.9830 - val_loss: 0.0425 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 461/1000\n",
      " - 3s - loss: 0.0493 - acc: 0.9848 - precision_m: 0.9857 - recall_m: 0.9829 - val_loss: 0.0417 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9917\n",
      "Epoch 462/1000\n",
      " - 3s - loss: 0.0512 - acc: 0.9838 - precision_m: 0.9859 - recall_m: 0.9824 - val_loss: 0.0439 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 463/1000\n",
      " - 3s - loss: 0.0493 - acc: 0.9833 - precision_m: 0.9854 - recall_m: 0.9828 - val_loss: 0.0500 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 464/1000\n",
      " - 3s - loss: 0.0461 - acc: 0.9864 - precision_m: 0.9885 - recall_m: 0.9846 - val_loss: 0.0425 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 465/1000\n",
      " - 3s - loss: 0.0439 - acc: 0.9874 - precision_m: 0.9887 - recall_m: 0.9863 - val_loss: 0.0503 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 466/1000\n",
      " - 3s - loss: 0.0473 - acc: 0.9869 - precision_m: 0.9883 - recall_m: 0.9854 - val_loss: 0.0559 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 467/1000\n",
      " - 3s - loss: 0.0499 - acc: 0.9847 - precision_m: 0.9862 - recall_m: 0.9827 - val_loss: 0.0500 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 468/1000\n",
      " - 3s - loss: 0.0440 - acc: 0.9868 - precision_m: 0.9879 - recall_m: 0.9851 - val_loss: 0.0366 - val_acc: 0.9934 - val_precision_m: 0.9969 - val_recall_m: 0.9937\n",
      "Epoch 469/1000\n",
      " - 3s - loss: 0.0449 - acc: 0.9877 - precision_m: 0.9887 - recall_m: 0.9865 - val_loss: 0.0489 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 470/1000\n",
      " - 3s - loss: 0.0479 - acc: 0.9857 - precision_m: 0.9870 - recall_m: 0.9849 - val_loss: 0.0365 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 471/1000\n",
      " - 3s - loss: 0.0445 - acc: 0.9860 - precision_m: 0.9871 - recall_m: 0.9849 - val_loss: 0.0443 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 472/1000\n",
      " - 3s - loss: 0.0444 - acc: 0.9857 - precision_m: 0.9871 - recall_m: 0.9840 - val_loss: 0.0470 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 473/1000\n",
      " - 3s - loss: 0.0429 - acc: 0.9865 - precision_m: 0.9883 - recall_m: 0.9854 - val_loss: 0.0422 - val_acc: 0.9945 - val_precision_m: 0.9948 - val_recall_m: 0.9948\n",
      "Epoch 474/1000\n",
      " - 3s - loss: 0.0430 - acc: 0.9864 - precision_m: 0.9877 - recall_m: 0.9850 - val_loss: 0.0441 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 475/1000\n",
      " - 3s - loss: 0.0493 - acc: 0.9859 - precision_m: 0.9869 - recall_m: 0.9846 - val_loss: 0.0423 - val_acc: 0.9934 - val_precision_m: 0.9948 - val_recall_m: 0.9937\n",
      "Epoch 476/1000\n",
      " - 3s - loss: 0.0480 - acc: 0.9849 - precision_m: 0.9864 - recall_m: 0.9836 - val_loss: 0.0548 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 477/1000\n",
      " - 3s - loss: 0.0438 - acc: 0.9871 - precision_m: 0.9886 - recall_m: 0.9856 - val_loss: 0.0391 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 478/1000\n",
      " - 3s - loss: 0.0446 - acc: 0.9858 - precision_m: 0.9873 - recall_m: 0.9849 - val_loss: 0.0509 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 479/1000\n",
      " - 3s - loss: 0.0504 - acc: 0.9858 - precision_m: 0.9871 - recall_m: 0.9843 - val_loss: 0.0463 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 480/1000\n",
      " - 3s - loss: 0.0400 - acc: 0.9895 - precision_m: 0.9905 - recall_m: 0.9875 - val_loss: 0.0449 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9906\n",
      "Epoch 481/1000\n",
      " - 3s - loss: 0.0436 - acc: 0.9859 - precision_m: 0.9869 - recall_m: 0.9849 - val_loss: 0.0397 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 482/1000\n",
      " - 3s - loss: 0.0473 - acc: 0.9868 - precision_m: 0.9880 - recall_m: 0.9856 - val_loss: 0.0443 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 483/1000\n",
      " - 3s - loss: 0.0408 - acc: 0.9870 - precision_m: 0.9884 - recall_m: 0.9857 - val_loss: 0.0419 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 484/1000\n",
      " - 3s - loss: 0.0399 - acc: 0.9884 - precision_m: 0.9895 - recall_m: 0.9869 - val_loss: 0.0392 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 485/1000\n",
      " - 3s - loss: 0.0403 - acc: 0.9871 - precision_m: 0.9879 - recall_m: 0.9854 - val_loss: 0.0377 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 486/1000\n",
      " - 3s - loss: 0.0438 - acc: 0.9866 - precision_m: 0.9876 - recall_m: 0.9852 - val_loss: 0.0433 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 487/1000\n",
      " - 3s - loss: 0.0419 - acc: 0.9877 - precision_m: 0.9890 - recall_m: 0.9866 - val_loss: 0.0426 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 488/1000\n",
      " - 3s - loss: 0.0437 - acc: 0.9877 - precision_m: 0.9884 - recall_m: 0.9866 - val_loss: 0.0368 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 489/1000\n",
      " - 3s - loss: 0.0494 - acc: 0.9848 - precision_m: 0.9865 - recall_m: 0.9830 - val_loss: 0.0380 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 490/1000\n",
      " - 3s - loss: 0.0408 - acc: 0.9880 - precision_m: 0.9892 - recall_m: 0.9868 - val_loss: 0.0452 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 491/1000\n",
      " - 3s - loss: 0.0412 - acc: 0.9881 - precision_m: 0.9896 - recall_m: 0.9860 - val_loss: 0.0372 - val_acc: 0.9967 - val_precision_m: 0.9969 - val_recall_m: 0.9969\n",
      "Epoch 492/1000\n",
      " - 3s - loss: 0.0485 - acc: 0.9847 - precision_m: 0.9863 - recall_m: 0.9833 - val_loss: 0.0380 - val_acc: 0.9967 - val_precision_m: 0.9989 - val_recall_m: 0.9969\n",
      "Epoch 493/1000\n",
      " - 3s - loss: 0.0436 - acc: 0.9881 - precision_m: 0.9889 - recall_m: 0.9874 - val_loss: 0.0362 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 494/1000\n",
      " - 3s - loss: 0.0532 - acc: 0.9832 - precision_m: 0.9843 - recall_m: 0.9817 - val_loss: 0.0391 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 495/1000\n",
      " - 3s - loss: 0.0446 - acc: 0.9857 - precision_m: 0.9868 - recall_m: 0.9839 - val_loss: 0.0460 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 496/1000\n",
      " - 3s - loss: 0.0387 - acc: 0.9877 - precision_m: 0.9886 - recall_m: 0.9863 - val_loss: 0.0457 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 497/1000\n",
      " - 3s - loss: 0.0394 - acc: 0.9884 - precision_m: 0.9895 - recall_m: 0.9872 - val_loss: 0.0429 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 498/1000\n",
      " - 3s - loss: 0.0384 - acc: 0.9885 - precision_m: 0.9895 - recall_m: 0.9882 - val_loss: 0.0434 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 499/1000\n",
      " - 3s - loss: 0.0382 - acc: 0.9887 - precision_m: 0.9899 - recall_m: 0.9879 - val_loss: 0.0499 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 500/1000\n",
      " - 3s - loss: 0.0450 - acc: 0.9863 - precision_m: 0.9871 - recall_m: 0.9849 - val_loss: 0.0448 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 501/1000\n",
      " - 3s - loss: 0.0461 - acc: 0.9857 - precision_m: 0.9872 - recall_m: 0.9839 - val_loss: 0.0379 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 502/1000\n",
      " - 3s - loss: 0.0392 - acc: 0.9882 - precision_m: 0.9900 - recall_m: 0.9874 - val_loss: 0.0510 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 503/1000\n",
      " - 3s - loss: 0.0428 - acc: 0.9880 - precision_m: 0.9891 - recall_m: 0.9869 - val_loss: 0.0402 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 504/1000\n",
      " - 3s - loss: 0.0418 - acc: 0.9868 - precision_m: 0.9879 - recall_m: 0.9850 - val_loss: 0.0481 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 505/1000\n",
      " - 3s - loss: 0.0410 - acc: 0.9868 - precision_m: 0.9881 - recall_m: 0.9855 - val_loss: 0.0320 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 506/1000\n",
      " - 3s - loss: 0.0366 - acc: 0.9884 - precision_m: 0.9895 - recall_m: 0.9869 - val_loss: 0.0450 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "Epoch 507/1000\n",
      " - 3s - loss: 0.0352 - acc: 0.9893 - precision_m: 0.9908 - recall_m: 0.9882 - val_loss: 0.0479 - val_acc: 0.9901 - val_precision_m: 0.9906 - val_recall_m: 0.9906\n",
      "Epoch 508/1000\n",
      " - 3s - loss: 0.0337 - acc: 0.9902 - precision_m: 0.9912 - recall_m: 0.9889 - val_loss: 0.0472 - val_acc: 0.9934 - val_precision_m: 0.9937 - val_recall_m: 0.9937\n",
      "\n",
      "Reached perfect accuracy so cancelling training!\n"
     ]
    }
   ],
   "source": [
    "# train the model\n",
    "# Here GPU fails. Before this line of code everything runs through i7 processor.\n",
    "bsize = 64\n",
    "history = model.fit(X_train, y_train, validation_data=(X_valid, y_valid), epochs=1000, batch_size = bsize, verbose=2, shuffle = True, callbacks = [epoch_schedule])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ibf8nlWdBNgH"
   },
   "source": [
    "## Plot Accuracy and Loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 474
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 60809,
     "status": "ok",
     "timestamp": 1571627063845,
     "user": {
      "displayName": "Mahindra Singh Rautela",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mBTSnpeHr1j42u0w7ABMmG3BYbWyCtFQVbOvr6sAA=s64",
      "userId": "15859880813264051870"
     },
     "user_tz": -330
    },
    "id": "kySYDbsQ7uON",
    "outputId": "4e673bae-c760-4b52-a4a1-0460bb47ab57"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#---Summarize history for loss\n",
    "plt.figure(figsize=(20,5))\n",
    "plt.plot(history.history['loss'],'-o')\n",
    "plt.plot(history.history['val_loss'],'-s')\n",
    "plt.title('Loss Curve',fontsize=20)\n",
    "plt.ylabel('Loss',fontsize=18)\n",
    "#plt.grid()\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.xlabel('Number of epochs',fontsize=18)\n",
    "plt.legend(['train', 'test'], loc='lower right',fontsize=18)\n",
    "#plt.axis([0,1000,0,500])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#---Summarize history for accuracy---\n",
    "plt.figure(figsize=(20,5))\n",
    "plt.plot(history.history['acc'],'-o')\n",
    "plt.plot(history.history['val_acc'],'-s')\n",
    "plt.title('Accuracy Curve',fontsize=20)\n",
    "plt.ylabel('Accuracy',fontsize=18)\n",
    "#plt.grid()\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.xlabel('Number of epochs',fontsize=18)\n",
    "plt.legend(['train', 'test'], loc='lower right',fontsize=18)\n",
    "#plt.axis([0,1000,0,500])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Output of final layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# trained model weights\n",
    "layer_outputs = model.layers[6].output\n",
    "print(layer_outputs.shape)\n",
    "print(layer_outputs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Weights and biases of any layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "layer_weights = model.layers[0].get_weights()[0]\n",
    "print(layer_weights.shape)\n",
    "print(layer_weights)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "layer_biases  = model.layers[0].get_weights()[1]\n",
    "print(layer_biases.shape)\n",
    "print(layer_biases)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.utils.multiclass import unique_labels\n",
    "import itertools\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.utils import to_categorical\n",
    "\n",
    "fig = plt.gcf()\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title=None,\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    if not title:\n",
    "        if normalize:\n",
    "            title = 'Normalized confusion matrix'\n",
    "        else:\n",
    "            title = 'Confusion matrix, without normalization'\n",
    "\n",
    "    plt.imshow(cm, interpolation = 'nearest', cmap = cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation = 45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "    \n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        cm = np.round(cm, 2)\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "    print(cm)\n",
    "\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "      plt.text(j, i, cm[i,j],\n",
    "      horizontalalignment = 'center',\n",
    "      color = \"white\" if cm[i,j] > thresh else \"black\")\n",
    "\n",
    "    fig.tight_layout()\n",
    "    plt.ylabel('True Label')\n",
    "    plt.xlabel('Predicted label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_predicted = model.predict(X_valid, verbose = 2)\n",
    "y_actual = y_valid\n",
    "\n",
    "y_predict = model.predict(X_valid, verbose = 2)\n",
    "y_predict = np.argmax(y_predict, axis=1)\n",
    "\n",
    "y_act = np.argmax(y_valid, axis=1)  # converting one hot representation back to numerical data\n",
    "\n",
    "cm = confusion_matrix(y_act, y_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[129   0   0   0   0   0   0]\n",
      " [  0 134   0   0   0   0   0]\n",
      " [  0   0 113   0   0   0   0]\n",
      " [  0   0   3 136   0   0   0]\n",
      " [  0   0   0   0 124   0   0]\n",
      " [  0   0   0   0   0 131   0]\n",
      " [  0   0   3   0   0   0 134]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot Non-Normalized confusion matrix\n",
    "cm_labels = ['DG1','DG2','DG3','DG4','DG5','DG6','DG7']\n",
    "plot_confusion_matrix(cm, classes=cm_labels, title='Confusion matrix')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normalized confusion matrix\n",
      "[[1.   0.   0.   0.   0.   0.   0.  ]\n",
      " [0.   1.   0.   0.   0.   0.   0.  ]\n",
      " [0.   0.   1.   0.   0.   0.   0.  ]\n",
      " [0.   0.   0.02 0.98 0.   0.   0.  ]\n",
      " [0.   0.   0.   0.   1.   0.   0.  ]\n",
      " [0.   0.   0.   0.   0.   1.   0.  ]\n",
      " [0.   0.   0.02 0.   0.   0.   0.98]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot Normalized confusion matrix\n",
    "plot_confusion_matrix(cm, classes=cm_labels, normalize = True, title='Confusion matrix')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         DG1       1.00      1.00      1.00       129\n",
      "         DG2       1.00      1.00      1.00       134\n",
      "         DG3       0.95      1.00      0.97       113\n",
      "         DG4       1.00      0.98      0.99       139\n",
      "         DG5       1.00      1.00      1.00       124\n",
      "         DG6       1.00      1.00      1.00       131\n",
      "         DG7       1.00      0.98      0.99       137\n",
      "\n",
      "    accuracy                           0.99       907\n",
      "   macro avg       0.99      0.99      0.99       907\n",
      "weighted avg       0.99      0.99      0.99       907\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Classification Report\n",
    "print('Classification Report')\n",
    "print(classification_report(y_act, y_predict, target_names = cm_labels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AUC-ROC curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compute ROC curve and ROC area for each class\n",
    "from sklearn.metrics import auc\n",
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "n_classes = 7\n",
    "fpr = dict()\n",
    "tpr = dict()\n",
    "roc_auc = dict()\n",
    "for i in range(n_classes):\n",
    "    fpr[i], tpr[i], thresholds = roc_curve(y_actual[:, i], np.round(y_predicted[:,i]))\n",
    "    roc_auc[i] = auc(fpr[i], tpr[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: array([0., 0., 1.]), 1: array([0., 0., 1.]), 2: array([0.        , 0.00755668, 1.        ]), 3: array([0., 0., 1.]), 4: array([0., 0., 1.]), 5: array([0., 0., 1.]), 6: array([0., 0., 1.])}\n"
     ]
    }
   ],
   "source": [
    "print(fpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: array([0., 1., 1.]), 1: array([0., 1., 1.]), 2: array([0., 1., 1.]), 3: array([0.        , 0.97841727, 1.        ]), 4: array([0., 1., 1.]), 5: array([0., 1., 1.]), 6: array([0.        , 0.97810219, 1.        ])}\n"
     ]
    }
   ],
   "source": [
    "print(tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: 1.0, 1: 1.0, 2: 0.9962216624685138, 3: 0.9892086330935252, 4: 1.0, 5: 1.0, 6: 0.9890510948905109}\n"
     ]
    }
   ],
   "source": [
    "print(roc_auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot of a ROC curve for all classes\n",
    "plt.figure(figsize=(8,8))\n",
    "for i in range(n_classes):\n",
    "    plt.plot(fpr[i], tpr[i], label='ROC curve (area = %0.2f)' % roc_auc[i])\n",
    "    plt.plot([0, 1], [0, 1], 'k--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xticks(fontsize=18)\n",
    "    plt.yticks(fontsize=18)\n",
    "    plt.xlabel('False Positive Rate',fontsize=18)\n",
    "    plt.ylabel('True Positive Rate',fontsize=18)\n",
    "    plt.title('Receiver operating characteristic example',fontsize=18)\n",
    "    plt.legend(cm_labels,loc=\"lower right\",fontsize=18)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "machine_shape": "hm",
   "name": "1DCNN_OGW_Binaryclassification.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}