CNN model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras.datasets import mnist\n",
    "from keras.layers import Dense, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.models import Sequential\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.layers import Activation\n",
    "from keras.layers import Dropout\n",
    "\n",
    "import matplotlib.pylab as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import sys\n",
    "import argparse\n",
    "import preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating dictionary of labels\n",
    "class_labels = {str(x):x for x in range(10)}\n",
    "class_labels.update({'+':10, 'times':11, '-':12 })\n",
    "label_class = dict( zip(class_labels.values(), class_labels.keys() ))\n",
    "\n",
    "# Loading data from .npy file and spliting into training and validation sets\n",
    "path = './Training Data/'\n",
    "ver = '1'\n",
    "path1 = path+'data_ver'+ver+'.npy'\n",
    "path2 = path+'labels_ver'+ver+'.npy'\n",
    "data, labels = preprocess.load_data(class_labels, path1 , path2, train = 0.9 , val = 0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvARkP1vxAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a1e742a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label =  4\n"
     ]
    }
   ],
   "source": [
    "idx = np.random.randint(data['train'].shape[0])\n",
    "plt.imshow(data['train'][idx,:].reshape(28,28),cmap = 'gray')\n",
    "plt.show()\n",
    "print('Label = ',label_class[ np.argmax(labels['train'][idx,:])])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 19543 samples, validate on 2171 samples\n",
      "Epoch 1/20\n",
      "19543/19543 [==============================] - 87s 4ms/step - loss: 0.2162 - acc: 0.9390 - val_loss: 5.3566 - val_acc: 0.1059\n",
      "Epoch 2/20\n",
      "19543/19543 [==============================] - 93s 5ms/step - loss: 0.0450 - acc: 0.9872 - val_loss: 6.3307 - val_acc: 0.1059\n",
      "Epoch 3/20\n",
      "19543/19543 [==============================] - 90s 5ms/step - loss: 0.0256 - acc: 0.9945 - val_loss: 6.9749 - val_acc: 0.1059\n",
      "Epoch 4/20\n",
      "19543/19543 [==============================] - 85s 4ms/step - loss: 0.0144 - acc: 0.9968 - val_loss: 6.5080 - val_acc: 0.1059\n",
      "Epoch 5/20\n",
      "19543/19543 [==============================] - 77s 4ms/step - loss: 0.0095 - acc: 0.9985 - val_loss: 4.6511 - val_acc: 0.2133\n",
      "Epoch 6/20\n",
      "19543/19543 [==============================] - 85s 4ms/step - loss: 0.0070 - acc: 0.9987 - val_loss: 2.1331 - val_acc: 0.4638\n",
      "Epoch 7/20\n",
      "19543/19543 [==============================] - 85s 4ms/step - loss: 0.0069 - acc: 0.9988 - val_loss: 0.6706 - val_acc: 0.7959\n",
      "Epoch 8/20\n",
      "19543/19543 [==============================] - 119s 6ms/step - loss: 0.0056 - acc: 0.9988 - val_loss: 0.1583 - val_acc: 0.9498\n",
      "Epoch 9/20\n",
      "19543/19543 [==============================] - 143s 7ms/step - loss: 0.0068 - acc: 0.9981 - val_loss: 0.0868 - val_acc: 0.9756\n",
      "Epoch 10/20\n",
      "19543/19543 [==============================] - 152s 8ms/step - loss: 0.0090 - acc: 0.9975 - val_loss: 0.0565 - val_acc: 0.9820\n",
      "Epoch 11/20\n",
      "19543/19543 [==============================] - 150s 8ms/step - loss: 0.0059 - acc: 0.9987 - val_loss: 0.0449 - val_acc: 0.9862\n",
      "Epoch 12/20\n",
      "19543/19543 [==============================] - 143s 7ms/step - loss: 0.0054 - acc: 0.9987 - val_loss: 0.0573 - val_acc: 0.9839\n",
      "Epoch 13/20\n",
      "19543/19543 [==============================] - 146s 7ms/step - loss: 0.0066 - acc: 0.9983 - val_loss: 0.0892 - val_acc: 0.9733\n",
      "Epoch 14/20\n",
      "19543/19543 [==============================] - 134s 7ms/step - loss: 0.0056 - acc: 0.9985 - val_loss: 0.0520 - val_acc: 0.9862\n",
      "Epoch 15/20\n",
      "19543/19543 [==============================] - 148s 8ms/step - loss: 0.0044 - acc: 0.9988 - val_loss: 0.0684 - val_acc: 0.9825\n",
      "Epoch 16/20\n",
      "19543/19543 [==============================] - 162s 8ms/step - loss: 0.0033 - acc: 0.9990 - val_loss: 0.0460 - val_acc: 0.9885\n",
      "Epoch 17/20\n",
      "19543/19543 [==============================] - 157s 8ms/step - loss: 0.0022 - acc: 0.9995 - val_loss: 0.0340 - val_acc: 0.9899\n",
      "Epoch 18/20\n",
      "19543/19543 [==============================] - 144s 7ms/step - loss: 0.0022 - acc: 0.9996 - val_loss: 0.0388 - val_acc: 0.9889\n",
      "Epoch 19/20\n",
      "19543/19543 [==============================] - 126s 6ms/step - loss: 0.0016 - acc: 0.9996 - val_loss: 0.0487 - val_acc: 0.9857\n",
      "Epoch 20/20\n",
      "19543/19543 [==============================] - 124s 6ms/step - loss: 0.0042 - acc: 0.9987 - val_loss: 0.0780 - val_acc: 0.9779\n",
      "Saved model to disk\n"
     ]
    }
   ],
   "source": [
    "img_x, img_y = 28, 28\n",
    "x_train = data['train'].reshape(data['train'].shape[0], img_x, img_y, 1)\n",
    "y_train = labels['train']\n",
    "x_test = data['val'].reshape(data['val'].shape[0], img_x, img_y, 1)\n",
    "y_test = labels['val']\n",
    "input_shape = (img_x, img_y, 1)\n",
    "\n",
    "# convert the data to the right type\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "num_classes = y_train.shape[1]\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, kernel_size=(3, 3), strides=(1, 1),input_shape=input_shape))\n",
    "model.add(BatchNormalization())\n",
    "model.add( keras.layers.LeakyReLU(alpha=0.1)  )\n",
    "model.add(Dropout(0.1))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "\n",
    "model.add(Conv2D(64, (3, 3)))\n",
    "model.add(BatchNormalization())\n",
    "model.add( keras.layers.LeakyReLU(alpha=0.1)  )\n",
    "model.add(Dropout(0.1))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(1000 ))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.1))\n",
    "\n",
    "model.add(Dense(256 ))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.1))\n",
    "\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "adam1 = keras.optimizers.Adam(lr=0.001, beta_1=0.9, beta_2=0.999)\n",
    "model.compile(loss=keras.losses.categorical_crossentropy, optimizer=adam1,metrics=['accuracy'] )\n",
    "history = model.fit(x_train, y_train, epochs=20, batch_size=256,validation_split=0.1,\n",
    "                   validation_data=(x_test, y_test))\n",
    "\n",
    "# serialize model to JSON\n",
    "model_json = model.to_json()\n",
    "with open(\"./trained models/CNN_ver{0}.1.json\".format(ver), \"w\") as json_file:\n",
    "    json_file.write(model_json)\n",
    "# serialize weights to HDF5\n",
    "model.save_weights(\"./trained models/CNN_ver{0}.1.h5\".format(ver))\n",
    "print(\"Saved model to disk\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "  CNN Ver1 results test accuracy = "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a3876d828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf4AAAEeCAYAAABrK3enAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a3c04b898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfEAAAEeCAYAAACaICwqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ni4iIFKj/D1TqjIsj40SXAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a25062f28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# summarize history for accuracy\n",
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('CNN model ver{0} accuracy'.format(ver),fontsize = 14)\n",
    "plt.xticks(fontsize = 13)\n",
    "plt.yticks(fontsize = 13)\n",
    "plt.ylabel('accuracy',fontsize = 14)\n",
    "plt.xlabel('epochs',fontsize = 14)\n",
    "plt.legend(['train', 'val'], loc='upper left',fontsize = 12)\n",
    "plt.show()\n",
    "\n",
    "# summarize history for loss\n",
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('CNN model ver{0} loss'.format(ver),fontsize = 14)\n",
    "plt.xticks(fontsize = 13)\n",
    "plt.yticks(fontsize = 13)\n",
    "plt.ylabel('loss',fontsize = 14)\n",
    "plt.xlabel('epochs',fontsize = 14)\n",
    "plt.legend(['train', 'val'], loc='upper right', fontsize = 12)\n",
    "plt.show()\n"
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}