model_trainig.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Face Emotion Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import Packages and define Learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import division, absolute_import\n",
    "\n",
    "from dataset_loader import DatasetLoader\n",
    "\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.models import Sequential, load_model\n",
    "from tensorflow.keras.layers import InputLayer, Conv2D, MaxPool2D, Flatten, Dense, Dropout, BatchNormalization\n",
    "\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, ReduceLROnPlateau\n",
    "\n",
    "\n",
    "from constants import *\n",
    "from os.path import isfile, join\n",
    "from os import makedirs\n",
    "from time import strftime, localtime\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.dummy import DummyClassifier\n",
    "\n",
    "import sys"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare Training and Test Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMPLETE DATA SET: images.shape: (14018, 48, 48, 1) labels.shape: (14018, 7)\n",
      "TRAINING DATA SET: images.shape: (11214, 48, 48, 1) labels.shape: (11214, 7)\n",
      "TESTIMG DATA SET: images.shape: (2804, 48, 48, 1) labels.shape: (2804, 7)\n"
     ]
    }
   ],
   "source": [
    "images = np.load(join('data', 'data_images.npy'))\n",
    "images = images.reshape([-1, SIZE_FACE, SIZE_FACE, 1])\n",
    "labels = np.load(join('data', 'data_labels.npy')).reshape([-1, len(EMOTIONS)])\n",
    "\n",
    "print('COMPLETE DATA SET: images.shape:',images.shape,'labels.shape:',labels.shape)\n",
    "train_x, test_x, train_y, test_y = train_test_split(images, labels, test_size=0.20, random_state=42)\n",
    "print('TRAINING DATA SET: images.shape:',train_x.shape,'labels.shape:',train_y.shape)\n",
    "print('TESTIMG DATA SET: images.shape:',test_x.shape,'labels.shape:',test_y.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Make Dummy Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.08238231098430814"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy = DummyClassifier(strategy='stratified')\n",
    "dummy.fit(X=train_x, y=train_y)\n",
    "dummy.score(X=test_x, y=test_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define Log Directory\n",
    "Start here to set a fresh log directory with."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "log_path = './logs/' + strftime('%Y-%m-%d-%H-%M',localtime())\n",
    "model_name = 'EmotionClassifier-EP{epoch:02d}-ACC{val_categorical_accuracy:.2f}.hdf5' \n",
    "\n",
    "makedirs(log_path, exist_ok=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define Keras Callbacks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "reduce_lr = ReduceLROnPlateau(monitor='val_categorical_accuracy', factor=0.1, verbose=1,\n",
    "                              patience=5, min_lr=1e-6)\n",
    "\n",
    "model_check_point = ModelCheckpoint(join(log_path,model_name), \n",
    "                                  monitor='val_categorical_accuracy',\n",
    "                                  verbose=1,\n",
    "                                  save_best_only=True,\n",
    "                                  mode='max')\n",
    "\n",
    "callbacks = [reduce_lr, model_check_point]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# No rescale, because already done!!\n",
    "\n",
    "image_data_gen_args = dict(#featurewise_center=False,\n",
    "                           #featurewise_std_normalization=False,\n",
    "                           rotation_range=30,\n",
    "                           width_shift_range=0.3,\n",
    "                           height_shift_range=0.3,\n",
    "                           horizontal_flip=False,\n",
    "                           zoom_range=0.3\n",
    "                          )\n",
    "\n",
    "datagen = ImageDataGenerator(image_data_gen_args)\n",
    "#datagen.fit(x=train_x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Definition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_22 (Conv2D)           (None, 46, 46, 32)        320       \n",
      "_________________________________________________________________\n",
      "conv2d_23 (Conv2D)           (None, 44, 44, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_10 (MaxPooling (None, 22, 22, 32)        0         \n",
      "_________________________________________________________________\n",
      "batch_normalization_v1_2 (Ba (None, 22, 22, 32)        128       \n",
      "_________________________________________________________________\n",
      "conv2d_24 (Conv2D)           (None, 22, 22, 64)        18496     \n",
      "_________________________________________________________________\n",
      "conv2d_25 (Conv2D)           (None, 22, 22, 64)        36928     \n",
      "_________________________________________________________________\n",
      "conv2d_26 (Conv2D)           (None, 22, 22, 64)        36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_11 (MaxPooling (None, 11, 11, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_27 (Conv2D)           (None, 9, 9, 128)         73856     \n",
      "_________________________________________________________________\n",
      "conv2d_28 (Conv2D)           (None, 4, 4, 128)         147584    \n",
      "_________________________________________________________________\n",
      "flatten_4 (Flatten)          (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 512)               1049088   \n",
      "_________________________________________________________________\n",
      "dense_9 (Dense)              (None, 7)                 3591      \n",
      "=================================================================\n",
      "Total params: 1,376,167\n",
      "Trainable params: 1,376,103\n",
      "Non-trainable params: 64\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "INPUT_SHAPE = [SIZE_FACE,SIZE_FACE,1]\n",
    "OUTPUT_SHAPE = len(EMOTIONS)\n",
    "\n",
    "\n",
    "vgg_like = Sequential([\n",
    "    #InputLayer(input_shape =(SIZE_FACE, SIZE_FACE, 1),name='input'),\n",
    "    Conv2D(filters=32,kernel_size=(3,3),input_shape=INPUT_SHAPE, padding='valid', activation='relu'),\n",
    "    Conv2D(filters=32,kernel_size=(3,3), padding='valid', activation='relu'),\n",
    "    MaxPool2D(pool_size=(3, 3), strides=2, padding='same'),\n",
    "    BatchNormalization(),\n",
    "    Conv2D(filters=64, kernel_size=(3,3), activation='relu', padding='same'),\n",
    "    Conv2D(filters=64, kernel_size=(3,3), activation='relu', padding='same'),\n",
    "    Conv2D(filters=64, kernel_size=(3,3), activation='relu', padding='same'),\n",
    "    MaxPool2D(pool_size=(3, 3), strides=2, padding='same'),\n",
    "    #BatchNormalization(),\n",
    "    Conv2D(filters=128,kernel_size=(3,3), activation='relu', padding='valid'),\n",
    "    Conv2D(filters=128,kernel_size=(3,3), activation='relu',strides=2, padding='valid'),    \n",
    "    #Conv2D(filters=128,kernel_size=(3,3), activation='relu', strides=2, padding='valid'),\n",
    "    #MaxPool2D(pool_size=(3, 3), strides=2, padding='same'),\n",
    "    Flatten(),\n",
    "    Dropout(0.5),\n",
    "    Dense(units=512,activation='relu'),\n",
    "    Dense(units=OUTPUT_SHAPE, activation='softmax')\n",
    "])\n",
    "\n",
    "\n",
    "\n",
    "model = Sequential([Conv2D(filters=32,kernel_size=(3,3),input_shape=INPUT_SHAPE, activation='relu'),\n",
    "                    MaxPool2D(pool_size=(3, 3), strides=2),\n",
    "                    Conv2D(filters=32,kernel_size=(3,3), activation='relu', padding='same'),\n",
    "                    MaxPool2D(pool_size=(3, 3), strides=2),\n",
    "                    Conv2D(filters=64, kernel_size=(3,3), activation='relu', padding='same'),\n",
    "                    MaxPool2D(pool_size=(3, 3), strides=2),\n",
    "                    Conv2D(filters=128,kernel_size=(3,3), activation='relu', strides=2, padding='same'),\n",
    "                    Flatten(),\n",
    "                    Dropout(0.5),\n",
    "                    Dense(units=512,activation='relu'),\n",
    "                    Dense(units=OUTPUT_SHAPE, activation='softmax')])\n",
    "\n",
    "model = vgg_like\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=Adam(lr=0.0001),\n",
    "              metrics=['categorical_accuracy'])\n",
    "\n",
    "# Save Generator Parameter and Model Summery in .txt file\n",
    "\n",
    "with open(join(log_path,'model_summery.txt'),'w+') as fh:\n",
    "    fh.write(str('ImageDataGenerator\\n'))\n",
    "    fh.write(str('==================\\n'))\n",
    "    fh.write(str(image_data_gen_args))\n",
    "    fh.write('\\n\\n\\n\\n')\n",
    "    # Pass the file handle in as a lambda function to make it callable\n",
    "    fh.write(str('Model Summery\\n'))\n",
    "    fh.write(str('=============\\n'))\n",
    "    model.summary(print_fn=lambda x: fh.write(x + '\\n'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Start Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "2804/2804 [==============================] - 1s 199us/sample - loss: 1.8851 - categorical_accuracy: 0.3516\n",
      "\n",
      "Epoch 00001: val_categorical_accuracy improved from -inf to 0.35164, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP01-ACC0.35.hdf5\n",
      "88/88 [==============================] - 5s 59ms/step - loss: 1.7465 - categorical_accuracy: 0.3281 - val_loss: 1.8851 - val_categorical_accuracy: 0.3516\n",
      "Epoch 2/50\n",
      "2804/2804 [==============================] - 0s 63us/sample - loss: 1.8030 - categorical_accuracy: 0.4793\n",
      "\n",
      "Epoch 00002: val_categorical_accuracy improved from 0.35164 to 0.47932, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP02-ACC0.48.hdf5\n",
      "88/88 [==============================] - 2s 21ms/step - loss: 1.4689 - categorical_accuracy: 0.4669 - val_loss: 1.8030 - val_categorical_accuracy: 0.4793\n",
      "Epoch 3/50\n",
      "2804/2804 [==============================] - 0s 55us/sample - loss: 1.6550 - categorical_accuracy: 0.5310\n",
      "\n",
      "Epoch 00003: val_categorical_accuracy improved from 0.47932 to 0.53103, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP03-ACC0.53.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 1.3278 - categorical_accuracy: 0.5149 - val_loss: 1.6551 - val_categorical_accuracy: 0.5310\n",
      "Epoch 4/50\n",
      "2804/2804 [==============================] - 0s 62us/sample - loss: 1.6229 - categorical_accuracy: 0.5082\n",
      "\n",
      "Epoch 00004: val_categorical_accuracy did not improve from 0.53103\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 1.2330 - categorical_accuracy: 0.5482 - val_loss: 1.6230 - val_categorical_accuracy: 0.5082\n",
      "Epoch 5/50\n",
      "2804/2804 [==============================] - 0s 61us/sample - loss: 1.4116 - categorical_accuracy: 0.5585\n",
      "\n",
      "Epoch 00005: val_categorical_accuracy improved from 0.53103 to 0.55849, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP05-ACC0.56.hdf5\n",
      "88/88 [==============================] - 2s 21ms/step - loss: 1.1753 - categorical_accuracy: 0.5726 - val_loss: 1.4118 - val_categorical_accuracy: 0.5585\n",
      "Epoch 6/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.2120 - categorical_accuracy: 0.5685\n",
      "\n",
      "Epoch 00006: val_categorical_accuracy improved from 0.55849 to 0.56847, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP06-ACC0.57.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 1.1159 - categorical_accuracy: 0.5921 - val_loss: 1.2122 - val_categorical_accuracy: 0.5685\n",
      "Epoch 7/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.1819 - categorical_accuracy: 0.5546\n",
      "\n",
      "Epoch 00007: val_categorical_accuracy did not improve from 0.56847\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 1.0833 - categorical_accuracy: 0.6047 - val_loss: 1.1821 - val_categorical_accuracy: 0.5546\n",
      "Epoch 8/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.0906 - categorical_accuracy: 0.5991\n",
      "\n",
      "Epoch 00008: val_categorical_accuracy improved from 0.56847 to 0.59914, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP08-ACC0.60.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 1.0376 - categorical_accuracy: 0.6199 - val_loss: 1.0905 - val_categorical_accuracy: 0.5991\n",
      "Epoch 9/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0725 - categorical_accuracy: 0.6016\n",
      "\n",
      "Epoch 00009: val_categorical_accuracy improved from 0.59914 to 0.60164, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP09-ACC0.60.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 1.0058 - categorical_accuracy: 0.6303 - val_loss: 1.0723 - val_categorical_accuracy: 0.6016\n",
      "Epoch 10/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.0666 - categorical_accuracy: 0.6009\n",
      "\n",
      "Epoch 00010: val_categorical_accuracy did not improve from 0.60164\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.9820 - categorical_accuracy: 0.6442 - val_loss: 1.0664 - val_categorical_accuracy: 0.6009\n",
      "Epoch 11/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0402 - categorical_accuracy: 0.6163\n",
      "\n",
      "Epoch 00011: val_categorical_accuracy improved from 0.60164 to 0.61626, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP11-ACC0.62.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.9537 - categorical_accuracy: 0.6542 - val_loss: 1.0400 - val_categorical_accuracy: 0.6163\n",
      "Epoch 12/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0290 - categorical_accuracy: 0.6213\n",
      "\n",
      "Epoch 00012: val_categorical_accuracy improved from 0.61626 to 0.62126, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP12-ACC0.62.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.9142 - categorical_accuracy: 0.6724 - val_loss: 1.0288 - val_categorical_accuracy: 0.6213\n",
      "Epoch 13/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0510 - categorical_accuracy: 0.6123\n",
      "\n",
      "Epoch 00013: val_categorical_accuracy did not improve from 0.62126\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.8842 - categorical_accuracy: 0.6838 - val_loss: 1.0508 - val_categorical_accuracy: 0.6123\n",
      "Epoch 14/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0225 - categorical_accuracy: 0.6198\n",
      "\n",
      "Epoch 00014: val_categorical_accuracy did not improve from 0.62126\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.8690 - categorical_accuracy: 0.6886 - val_loss: 1.0223 - val_categorical_accuracy: 0.6198\n",
      "Epoch 15/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0326 - categorical_accuracy: 0.6273\n",
      "\n",
      "Epoch 00015: val_categorical_accuracy improved from 0.62126 to 0.62732, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP15-ACC0.63.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.8391 - categorical_accuracy: 0.6979 - val_loss: 1.0324 - val_categorical_accuracy: 0.6273\n",
      "Epoch 16/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.0064 - categorical_accuracy: 0.6327\n",
      "\n",
      "Epoch 00016: val_categorical_accuracy improved from 0.62732 to 0.63267, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP16-ACC0.63.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.8260 - categorical_accuracy: 0.7012 - val_loss: 1.0060 - val_categorical_accuracy: 0.6327\n",
      "Epoch 17/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0144 - categorical_accuracy: 0.6259\n",
      "\n",
      "Epoch 00017: val_categorical_accuracy did not improve from 0.63267\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.7918 - categorical_accuracy: 0.7141 - val_loss: 1.0143 - val_categorical_accuracy: 0.6259\n",
      "Epoch 18/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0131 - categorical_accuracy: 0.6427\n",
      "\n",
      "Epoch 00018: val_categorical_accuracy improved from 0.63267 to 0.64265, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP18-ACC0.64.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.7715 - categorical_accuracy: 0.7254 - val_loss: 1.0129 - val_categorical_accuracy: 0.6427\n",
      "Epoch 19/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 0.9989 - categorical_accuracy: 0.6402\n",
      "\n",
      "Epoch 00019: val_categorical_accuracy did not improve from 0.64265\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.7478 - categorical_accuracy: 0.7312 - val_loss: 0.9985 - val_categorical_accuracy: 0.6402\n",
      "Epoch 20/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0293 - categorical_accuracy: 0.6266\n",
      "\n",
      "Epoch 00020: val_categorical_accuracy did not improve from 0.64265\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.7247 - categorical_accuracy: 0.7404 - val_loss: 1.0290 - val_categorical_accuracy: 0.6266\n",
      "Epoch 21/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0142 - categorical_accuracy: 0.6412\n",
      "\n",
      "Epoch 00021: val_categorical_accuracy did not improve from 0.64265\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.7026 - categorical_accuracy: 0.7475 - val_loss: 1.0139 - val_categorical_accuracy: 0.6412\n",
      "Epoch 22/50\n",
      "2804/2804 [==============================] - 0s 57us/sample - loss: 1.0091 - categorical_accuracy: 0.6473\n",
      "\n",
      "Epoch 00022: val_categorical_accuracy improved from 0.64265 to 0.64729, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP22-ACC0.65.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.6742 - categorical_accuracy: 0.7591 - val_loss: 1.0085 - val_categorical_accuracy: 0.6473\n",
      "Epoch 23/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.0135 - categorical_accuracy: 0.6476\n",
      "\n",
      "Epoch 00023: val_categorical_accuracy improved from 0.64729 to 0.64765, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP23-ACC0.65.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.6548 - categorical_accuracy: 0.7654 - val_loss: 1.0133 - val_categorical_accuracy: 0.6476\n",
      "Epoch 24/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.0016 - categorical_accuracy: 0.6476\n",
      "\n",
      "Epoch 00024: val_categorical_accuracy did not improve from 0.64765\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.6241 - categorical_accuracy: 0.7788 - val_loss: 1.0013 - val_categorical_accuracy: 0.6476\n",
      "Epoch 25/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0263 - categorical_accuracy: 0.6427\n",
      "\n",
      "Epoch 00025: val_categorical_accuracy did not improve from 0.64765\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.6027 - categorical_accuracy: 0.7812 - val_loss: 1.0261 - val_categorical_accuracy: 0.6427\n",
      "Epoch 26/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0254 - categorical_accuracy: 0.6548\n",
      "\n",
      "Epoch 00026: val_categorical_accuracy improved from 0.64765 to 0.65478, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP26-ACC0.65.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.5840 - categorical_accuracy: 0.7932 - val_loss: 1.0250 - val_categorical_accuracy: 0.6548\n",
      "Epoch 27/50\n",
      "2804/2804 [==============================] - 0s 57us/sample - loss: 1.0399 - categorical_accuracy: 0.6359\n",
      "\n",
      "Epoch 00027: val_categorical_accuracy did not improve from 0.65478\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.5694 - categorical_accuracy: 0.7967 - val_loss: 1.0396 - val_categorical_accuracy: 0.6359\n",
      "Epoch 28/50\n",
      "2804/2804 [==============================] - 0s 55us/sample - loss: 1.0007 - categorical_accuracy: 0.6562\n",
      "\n",
      "Epoch 00028: val_categorical_accuracy improved from 0.65478 to 0.65621, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP28-ACC0.66.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.5569 - categorical_accuracy: 0.7984 - val_loss: 1.0005 - val_categorical_accuracy: 0.6562\n",
      "Epoch 29/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0456 - categorical_accuracy: 0.6312\n",
      "\n",
      "Epoch 00029: val_categorical_accuracy did not improve from 0.65621\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.5292 - categorical_accuracy: 0.8173 - val_loss: 1.0452 - val_categorical_accuracy: 0.6312\n",
      "Epoch 30/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0152 - categorical_accuracy: 0.6583\n",
      "\n",
      "Epoch 00030: val_categorical_accuracy improved from 0.65621 to 0.65835, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP30-ACC0.66.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.5116 - categorical_accuracy: 0.8192 - val_loss: 1.0148 - val_categorical_accuracy: 0.6583\n",
      "Epoch 31/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0795 - categorical_accuracy: 0.6380\n",
      "\n",
      "Epoch 00031: val_categorical_accuracy did not improve from 0.65835\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.4821 - categorical_accuracy: 0.8263 - val_loss: 1.0792 - val_categorical_accuracy: 0.6380\n",
      "Epoch 32/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.0244 - categorical_accuracy: 0.6594\n",
      "\n",
      "Epoch 00032: val_categorical_accuracy improved from 0.65835 to 0.65942, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP32-ACC0.66.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.4673 - categorical_accuracy: 0.8345 - val_loss: 1.0242 - val_categorical_accuracy: 0.6594\n",
      "Epoch 33/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.0593 - categorical_accuracy: 0.6576\n",
      "\n",
      "Epoch 00033: val_categorical_accuracy did not improve from 0.65942\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.4634 - categorical_accuracy: 0.8355 - val_loss: 1.0591 - val_categorical_accuracy: 0.6576\n",
      "Epoch 34/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0662 - categorical_accuracy: 0.6562\n",
      "\n",
      "Epoch 00034: val_categorical_accuracy did not improve from 0.65942\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.4380 - categorical_accuracy: 0.8480 - val_loss: 1.0658 - val_categorical_accuracy: 0.6562\n",
      "Epoch 35/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0690 - categorical_accuracy: 0.6605\n",
      "\n",
      "Epoch 00035: val_categorical_accuracy improved from 0.65942 to 0.66049, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP35-ACC0.66.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.4150 - categorical_accuracy: 0.8546 - val_loss: 1.0685 - val_categorical_accuracy: 0.6605\n",
      "Epoch 36/50\n",
      "2804/2804 [==============================] - 0s 55us/sample - loss: 1.1145 - categorical_accuracy: 0.6608\n",
      "\n",
      "Epoch 00036: val_categorical_accuracy improved from 0.66049 to 0.66084, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP36-ACC0.66.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.3974 - categorical_accuracy: 0.8571 - val_loss: 1.1141 - val_categorical_accuracy: 0.6608\n",
      "Epoch 37/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.1280 - categorical_accuracy: 0.6345\n",
      "\n",
      "Epoch 00037: val_categorical_accuracy did not improve from 0.66084\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.3845 - categorical_accuracy: 0.8646 - val_loss: 1.1279 - val_categorical_accuracy: 0.6345\n",
      "Epoch 38/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.0635 - categorical_accuracy: 0.6619\n",
      "\n",
      "Epoch 00038: val_categorical_accuracy improved from 0.66084 to 0.66191, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP38-ACC0.66.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.3692 - categorical_accuracy: 0.8692 - val_loss: 1.0631 - val_categorical_accuracy: 0.6619\n",
      "Epoch 39/50\n",
      "2804/2804 [==============================] - 0s 60us/sample - loss: 1.1103 - categorical_accuracy: 0.6669\n",
      "\n",
      "Epoch 00039: val_categorical_accuracy improved from 0.66191 to 0.66690, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP39-ACC0.67.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.3527 - categorical_accuracy: 0.8739 - val_loss: 1.1099 - val_categorical_accuracy: 0.6669\n",
      "Epoch 40/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.0950 - categorical_accuracy: 0.6708\n",
      "\n",
      "Epoch 00040: val_categorical_accuracy improved from 0.66690 to 0.67083, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP40-ACC0.67.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.3382 - categorical_accuracy: 0.8822 - val_loss: 1.0948 - val_categorical_accuracy: 0.6708\n",
      "Epoch 41/50\n",
      "2804/2804 [==============================] - 0s 56us/sample - loss: 1.1293 - categorical_accuracy: 0.6637\n",
      "\n",
      "Epoch 00041: val_categorical_accuracy did not improve from 0.67083\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.3218 - categorical_accuracy: 0.8849 - val_loss: 1.1291 - val_categorical_accuracy: 0.6637\n",
      "Epoch 42/50\n",
      "2804/2804 [==============================] - 0s 58us/sample - loss: 1.1661 - categorical_accuracy: 0.6612\n",
      "\n",
      "Epoch 00042: val_categorical_accuracy did not improve from 0.67083\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.3086 - categorical_accuracy: 0.8933 - val_loss: 1.1658 - val_categorical_accuracy: 0.6612\n",
      "Epoch 43/50\n",
      "2804/2804 [==============================] - 0s 56us/sample - loss: 1.1479 - categorical_accuracy: 0.6619\n",
      "\n",
      "Epoch 00043: val_categorical_accuracy did not improve from 0.67083\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.2980 - categorical_accuracy: 0.8962 - val_loss: 1.1477 - val_categorical_accuracy: 0.6619\n",
      "Epoch 44/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.1417 - categorical_accuracy: 0.6669\n",
      "\n",
      "Epoch 00044: val_categorical_accuracy did not improve from 0.67083\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.2836 - categorical_accuracy: 0.8995 - val_loss: 1.1414 - val_categorical_accuracy: 0.6669\n",
      "Epoch 45/50\n",
      "2804/2804 [==============================] - 0s 58us/sample - loss: 1.1858 - categorical_accuracy: 0.6619\n",
      "\n",
      "Epoch 00045: ReduceLROnPlateau reducing learning rate to 9.999999747378752e-06.\n",
      "\n",
      "Epoch 00045: val_categorical_accuracy did not improve from 0.67083\n",
      "88/88 [==============================] - 2s 22ms/step - loss: 0.2713 - categorical_accuracy: 0.9060 - val_loss: 1.1851 - val_categorical_accuracy: 0.6619\n",
      "Epoch 46/50\n",
      "2804/2804 [==============================] - 0s 55us/sample - loss: 1.1896 - categorical_accuracy: 0.6715\n",
      "\n",
      "Epoch 00046: val_categorical_accuracy improved from 0.67083 to 0.67154, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP46-ACC0.67.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.2177 - categorical_accuracy: 0.9269 - val_loss: 1.1893 - val_categorical_accuracy: 0.6715\n",
      "Epoch 47/50\n",
      "2804/2804 [==============================] - 0s 54us/sample - loss: 1.1956 - categorical_accuracy: 0.6705\n",
      "\n",
      "Epoch 00047: val_categorical_accuracy did not improve from 0.67154\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.2059 - categorical_accuracy: 0.9316 - val_loss: 1.1953 - val_categorical_accuracy: 0.6705\n",
      "Epoch 48/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.1960 - categorical_accuracy: 0.6730\n",
      "\n",
      "Epoch 00048: val_categorical_accuracy improved from 0.67154 to 0.67297, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP48-ACC0.67.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.2076 - categorical_accuracy: 0.9294 - val_loss: 1.1956 - val_categorical_accuracy: 0.6730\n",
      "Epoch 49/50\n",
      "2804/2804 [==============================] - 0s 53us/sample - loss: 1.2023 - categorical_accuracy: 0.6708\n",
      "\n",
      "Epoch 00049: val_categorical_accuracy did not improve from 0.67297\n",
      "88/88 [==============================] - 2s 19ms/step - loss: 0.2027 - categorical_accuracy: 0.9353 - val_loss: 1.2020 - val_categorical_accuracy: 0.6708\n",
      "Epoch 50/50\n",
      "2804/2804 [==============================] - 0s 52us/sample - loss: 1.2050 - categorical_accuracy: 0.6744\n",
      "\n",
      "Epoch 00050: val_categorical_accuracy improved from 0.67297 to 0.67439, saving model to ./logs/2019-04-04-08-51/EmotionClassifier-EP50-ACC0.67.hdf5\n",
      "88/88 [==============================] - 2s 20ms/step - loss: 0.1987 - categorical_accuracy: 0.9362 - val_loss: 1.2047 - val_categorical_accuracy: 0.6744\n"
     ]
    }
   ],
   "source": [
    "BATCH_SIZE = 128\n",
    "EPOCH = 50\n",
    "\n",
    "history = model.fit_generator(\n",
    "    generator=datagen.flow(x=train_x,y=train_y,batch_size=BATCH_SIZE),\n",
    "    steps_per_epoch=2*(len(train_x)/BATCH_SIZE),\n",
    "    epochs=EPOCH,\n",
    "    #validation_data=test_datagen.flow(x=test_x,y=test_y,batch_size=BATCH_SIZE),\n",
    "    validation_data=(test_x,test_y),\n",
    "    #validation_steps=len(test_x)/BATCH_SIZE,\n",
    "    callbacks=callbacks\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modell Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(20,5))\n",
    "fig.suptitle('Model History',fontsize=16)\n",
    "\n",
    "# summarize history for loss\n",
    "ax = fig.add_subplot(121)\n",
    "ax.plot(history.history['categorical_accuracy'])\n",
    "ax.plot(history.history['val_categorical_accuracy'])\n",
    "# summarize history for accuracy\n",
    "ax.set_title('model accuracy')\n",
    "ax.set_ylabel('accuracy')\n",
    "ax.set_yticks(np.arange(0,1.1,step=.1))\n",
    "ax.set_xlabel('epoch')\n",
    "ax.legend(['train', 'test'])#, loc='upper left')\n",
    "\n",
    "# summarize history for loss\n",
    "ax = fig.add_subplot(122)\n",
    "ax.plot(history.history['loss'])\n",
    "ax.plot(history.history['val_loss'])\n",
    "ax.set_title('model loss')\n",
    "ax.set_ylabel('loss')\n",
    "ax.set_yticks(np.arange(0,2.6,step=.25))\n",
    "ax.set_xlabel('epoch')\n",
    "ax.legend(['train', 'test'])#, loc='upper left')\n",
    "plt.savefig(join(log_path,'model_history.png'), format='png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Confusion Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2804/2804 [==============================] - 0s 139us/step\n",
      "[1.064033034694688, 0.61269614835948649]\n",
      "2804/2804 [==============================] - 0s 169us/step\n",
      "[1.281970519312098, 0.67546362331012177]\n"
     ]
    }
   ],
   "source": [
    "top_acc_model = load_model(filepath='./logs/2019-03-11-10-11/EmotionClassifier-EP35-ACC0.68.hdf5')\n",
    "current_model = load_model(filepath='./data/emotion_recognition_model.hdf5')\n",
    "\n",
    "print(current_model.evaluate(x=test_x, y=test_y))\n",
    "print(top_acc_model.evaluate(x=test_x, y=test_y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[3 5 6 ..., 6 4 6]\n",
      "[3 5 6 ..., 6 4 6]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[140,   1,  21,  21,  49,  12,  79],\n",
       "       [ 20,  10,   1,   5,   7,   2,   8],\n",
       "       [ 57,   2,  84,  36,  53,  57,  83],\n",
       "       [ 10,   0,  15, 698,  12,  22,  60],\n",
       "       [ 30,   1,  19,  34, 120,   7, 138],\n",
       "       [ 11,   0,  12,  23,   6, 230,  28],\n",
       "       [ 21,   0,  10,  57,  48,   8, 436]])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction = current_model.predict(x=test_x)\n",
    "prediction = np.argmax(prediction,axis=1)\n",
    "true_class = np.argmax(test_y, axis=1)\n",
    "print(prediction)\n",
    "print(true_class)\n",
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "confusion_matrix(y_true=true_class,y_pred=prediction)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.utils.multiclass import unique_labels\n",
    "\n",
    "def plot_confusion_matrix(y_true, y_pred, 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",
    "    # Compute confusion matrix\n",
    "    cm = confusion_matrix(y_true, y_pred)\n",
    "    # Only use the labels that appear in the data\n",
    "    ulist = unique_labels(y_true, y_pred)\n",
    "    #print(classes[ulist])\n",
    "    #classes = classes[unique_labels(y_true, y_pred)]\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        #print(\"Normalized confusion matrix\")\n",
    "    #else:\n",
    "        #print('Confusion matrix, without normalization')\n",
    "\n",
    "    #print(cm)\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "    im = ax.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    ax.figure.colorbar(im, ax=ax)\n",
    "    # We want to show all ticks...\n",
    "    ax.set(xticks=np.arange(cm.shape[1]),\n",
    "           yticks=np.arange(cm.shape[0]),\n",
    "           # ... and label them with the respective list entries\n",
    "           xticklabels=classes, yticklabels=classes,\n",
    "           title=title,\n",
    "           ylabel='True label',\n",
    "           xlabel='Predicted label')\n",
    "\n",
    "    # Rotate the tick labels and set their alignment.\n",
    "    plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\",\n",
    "             rotation_mode=\"anchor\")\n",
    "\n",
    "    # Loop over data dimensions and create text annotations.\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i in range(cm.shape[0]):\n",
    "        for j in range(cm.shape[1]):\n",
    "            ax.text(j, i, format(cm[i, j], fmt),\n",
    "                    ha=\"center\", va=\"center\",\n",
    "                    color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "    fig.tight_layout()\n",
    "    return ax"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "np.set_printoptions(precision=2)\n",
    "\n",
    "# Plot non-normalized confusion matrix\n",
    "plot_confusion_matrix(prediction, true_class, classes=EMOTIONS,\n",
    "                      title='Confusion matrix, without normalization')\n",
    "\n",
    "# Plot normalized confusion matrix\n",
    "plot_confusion_matrix(prediction, true_class, classes=EMOTIONS, normalize=True,\n",
    "                      title='Normalized confusion matrix')\n",
    "\n",
    "#plt.savefig(join(log_path,'confusion_matrix.png'), format='png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "np.set_printoptions(precision=2)\n",
    "\n",
    "# Plot non-normalized confusion matrix\n",
    "plot_confusion_matrix(prediction, true_class, classes=EMOTIONS,\n",
    "                      title='Confusion matrix, without normalization')\n",
    "\n",
    "# Plot normalized confusion matrix\n",
    "plot_confusion_matrix(prediction, true_class, classes=EMOTIONS, normalize=True,\n",
    "                      title='Normalized confusion matrix')\n",
    "\n",
    "#plt.savefig(join(log_path,'confusion_matrix.png'), format='png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}