Model-checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = \"./train/\"\n",
    "validation = \"./validation/\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Details of Train Data : \n",
      "Total img in Angry dir is : 3995\n",
      "Total img in Disgust dir is : 436\n",
      "Total img in Fear dir is : 4097\n",
      "Total img in Happy dir is : 7215\n",
      "Total img in Neutral dir is : 4965\n",
      "Total img in Sad dir is : 4830\n",
      "Total img in Surprise dir is : 3171\n"
     ]
    }
   ],
   "source": [
    "print(\"Details of Train Data : \")\n",
    "for i in os.listdir(train):\n",
    "    print(\"Total img in %s dir is : %d\"%(i,len(os.listdir(train+i+\"/\"))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Details of Validation Data : \n",
      "Total img in Angry dir is : 491\n",
      "Total img in Disgust dir is : 55\n",
      "Total img in Fear dir is : 528\n",
      "Total img in Happy dir is : 879\n",
      "Total img in Neutral dir is : 626\n",
      "Total img in Sad dir is : 594\n",
      "Total img in Surprise dir is : 416\n"
     ]
    }
   ],
   "source": [
    "print(\"Details of Validation Data : \")\n",
    "for i in os.listdir(validation):\n",
    "    print(\"Total img in %s dir is : %d\"%(i,len(os.listdir(validation+i+\"/\"))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dir = {}\n",
    "for i in os.listdir(train):\n",
    "    train_dir[i] = train + i + \"/\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(48, 48, 3)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(48, 48, 3)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(48, 48, 3)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in os.listdir(train_dir['Happy'])[:3]:\n",
    "    img = cv2.imread(train_dir['Happy']+i)\n",
    "    print(img.shape)\n",
    "    plt.imshow(img)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_datagen = ImageDataGenerator(rescale=1.0/255.0 ,\n",
    "                                   rotation_range=10,\n",
    "                                   zoom_range=0.3,\n",
    "                                   shear_range=0.3,\n",
    "                                   horizontal_flip=True,\n",
    "                                   width_shift_range=0.4,\n",
    "                                   height_shift_range=0.4,fill_mode='nearest')\n",
    "validation_datagen = ImageDataGenerator(rescale=1.0/255.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 28709 images belonging to 7 classes.\n",
      "Found 3589 images belonging to 7 classes.\n"
     ]
    }
   ],
   "source": [
    "train_data = train_datagen.flow_from_directory(train,target_size=(50,50),class_mode='categorical',batch_size=32,color_mode='grayscale')\n",
    "validation_data = validation_datagen.flow_from_directory(validation,target_size=(50,50),class_mode='categorical',batch_size=32,color_mode='grayscale')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "model1 = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.Conv2D(32,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal',input_shape=(48,48,1)),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.Conv2D(32,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.MaxPool2D((2,2)),\n",
    "    tf.keras.layers.Dropout(0.2),\n",
    "    \n",
    "    tf.keras.layers.Conv2D(64,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.Conv2D(64,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.MaxPool2D((2,2)),\n",
    "    tf.keras.layers.Dropout(0.2),\n",
    "    \n",
    "    tf.keras.layers.Conv2D(128,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.Conv2D(128,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.MaxPool2D((2,2)),\n",
    "    tf.keras.layers.Dropout(0.2),\n",
    "    \n",
    "    tf.keras.layers.Conv2D(256,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.Conv2D(256,(3,3),activation='elu',padding=\"same\",kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.MaxPool2D((2,2)),\n",
    "    tf.keras.layers.Dropout(0.2),\n",
    "    \n",
    "    tf.keras.layers.Flatten(),\n",
    "    tf.keras.layers.Dense(64,activation='elu',kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.Dropout(0.5),\n",
    "    \n",
    "    tf.keras.layers.Dense(64,activation='elu',kernel_initializer='he_normal'),\n",
    "    tf.keras.layers.BatchNormalization(),\n",
    "    tf.keras.layers.Dropout(0.5),\n",
    "    \n",
    "    tf.keras.layers.Dense(7,activation='softmax',kernel_initializer='he_normal')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "#model1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.callbacks import ModelCheckpoint,EarlyStopping,ReduceLROnPlateau"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "checkpoint = ModelCheckpoint(\"Emotion.h5\",\n",
    "                            monitor='val_loss',\n",
    "                            mode='min',\n",
    "                            save_best_only=True,\n",
    "                            verbose=1)\n",
    "\n",
    "earlystop = EarlyStopping(monitor='val_loss',\n",
    "                         min_delta=0,\n",
    "                         patience=9,\n",
    "                         verbose=2,\n",
    "                         restore_best_weights=True)\n",
    "\n",
    "reduce_lr = ReduceLROnPlateau(monitor='val_loss',\n",
    "                             factor=0.2,\n",
    "                             patience=3,\n",
    "                             verbose=1,\n",
    "                             min_delta=0.0001)\n",
    "callbacks = [earlystop,checkpoint,reduce_lr]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "model1.compile(optimizer=tf.keras.optimizers.Adam(lr=0.001),loss=tf.keras.losses.categorical_crossentropy,metrics=['acc'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:sample_weight modes were coerced from\n",
      "  ...\n",
      "    to  \n",
      "  ['...']\n",
      "WARNING:tensorflow:sample_weight modes were coerced from\n",
      "  ...\n",
      "    to  \n",
      "  ['...']\n",
      "Train for 897 steps, validate for 112 steps\n",
      "Epoch 1/25\n"
     ]
    }
   ],
   "source": [
    "history = model1.fit(train_data,\n",
    "                    steps_per_epoch=train_data.n//train_data.batch_size,\n",
    "                    epochs=25,\n",
    "                    callbacks=callbacks,\n",
    "                    validation_data=validation_data,\n",
    "                    validation_steps=validation_data.n//validation_data.batch_size)"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}