model1-Conv_2-batchnormalisation-augmentation-2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "import keras\n",
    "from keras.callbacks import *\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.models import Sequential\n",
    "from keras.layers import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "traindata = pd.read_csv('sign_mnist_train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainlabel=traindata['label'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "traindata.drop('label',inplace=True,axis=1)\n",
    "trainimages = traindata.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainimages=trainimages.reshape(-1,28,28,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "testdata = pd.read_csv('sign_mnist_test.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "testlabel=testdata['label'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "testdata.drop('label',inplace=True,axis=1)\n",
    "testimages = testdata.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "testimages=testimages.reshape(-1,28,28,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "traingen=ImageDataGenerator(rotation_range=20,zoom_range=0.1,width_shift_range=0.1,height_shift_range=0.1,\n",
    "                  shear_range=0.1,horizontal_flip=True,rescale=1/255.0,validation_split=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "traindata_generator = traingen.flow(trainimages,trainlabel,subset='training')\n",
    "validationdata_generator = traingen.flow(trainimages,trainlabel,subset='validation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "testgen=ImageDataGenerator(rescale=1/255.0)\n",
    "testdata_generator = testgen.flow(testimages,testlabel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "model=Sequential([])\n",
    "\n",
    "model.add(Conv2D(64,(3,3),activation=\"relu\",input_shape=(28,28,1)))\n",
    "model.add(BatchNormalization())\n",
    "# model.add(Conv2D(64,(3,3),activation=\"relu\"))\n",
    "# model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(2,2))\n",
    "#model.add(Dropout(0.4))\n",
    "\n",
    "model.add(Conv2D(128,(3,3),activation=\"relu\"))\n",
    "model.add(BatchNormalization())\n",
    "# model.add(Conv2D(128,(3,3),activation=\"relu\"))\n",
    "# model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(2,2))\n",
    "#model.add(Dropout(0.4))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(BatchNormalization())\n",
    "#model.add(Dropout(0.4))\n",
    "model.add(Dense(256,activation=\"relu\"))\n",
    "model.add(BatchNormalization())\n",
    "#model.add(Dropout(0.4))\n",
    "model.add(Dense(26,activation=\"softmax\"))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss=\"sparse_categorical_crossentropy\",optimizer='adam',metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 26, 26, 64)        640       \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 26, 26, 64)        256       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 13, 13, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 11, 11, 128)       73856     \n",
      "_________________________________________________________________\n",
      "batch_normalization_2 (Batch (None, 11, 11, 128)       512       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 5, 5, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 3200)              0         \n",
      "_________________________________________________________________\n",
      "batch_normalization_3 (Batch (None, 3200)              12800     \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 256)               819456    \n",
      "_________________________________________________________________\n",
      "batch_normalization_4 (Batch (None, 256)               1024      \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 26)                6682      \n",
      "=================================================================\n",
      "Total params: 915,226\n",
      "Trainable params: 907,930\n",
      "Non-trainable params: 7,296\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define a Callback class that stops training once accuracy reaches 97.0%\n",
    "class myCallback(Callback):\n",
    "  def on_epoch_end(self, epoch, logs={}):\n",
    "    if(logs.get('val_accuracy')>0.995):\n",
    "      print(\"\\nReached 99.5% accuracy so cancelling training!\")\n",
    "      self.model.stop_training = True\n",
    "callback=myCallback()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.6587 - accuracy: 0.7915 - val_loss: 0.4311 - val_accuracy: 0.8211\n",
      "Epoch 2/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.2813 - accuracy: 0.9117 - val_loss: 0.4937 - val_accuracy: 0.9071\n",
      "Epoch 3/50\n",
      "773/773 [==============================] - 8s 11ms/step - loss: 0.1893 - accuracy: 0.9406 - val_loss: 1.1174 - val_accuracy: 0.7104\n",
      "Epoch 4/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.1564 - accuracy: 0.9494 - val_loss: 6.7603 - val_accuracy: 0.4302\n",
      "Epoch 5/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.1248 - accuracy: 0.9607 - val_loss: 1.7861 - val_accuracy: 0.5439\n",
      "Epoch 6/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.1189 - accuracy: 0.9616 - val_loss: 1.4909 - val_accuracy: 0.5865\n",
      "Epoch 7/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.1088 - accuracy: 0.9642 - val_loss: 0.8824 - val_accuracy: 0.6189\n",
      "Epoch 8/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.1102 - accuracy: 0.9633 - val_loss: 151.4670 - val_accuracy: 0.0503\n",
      "Epoch 9/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0979 - accuracy: 0.9675 - val_loss: 0.1491 - val_accuracy: 0.9475\n",
      "Epoch 10/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0799 - accuracy: 0.9749 - val_loss: 0.1940 - val_accuracy: 0.9738\n",
      "Epoch 11/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0657 - accuracy: 0.9794 - val_loss: 0.2236 - val_accuracy: 0.8186\n",
      "Epoch 12/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0680 - accuracy: 0.9786 - val_loss: 0.2963 - val_accuracy: 0.9097\n",
      "Epoch 13/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0688 - accuracy: 0.9769 - val_loss: 740.2717 - val_accuracy: 0.0765\n",
      "Epoch 14/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0743 - accuracy: 0.9748 - val_loss: 0.2507 - val_accuracy: 0.8175\n",
      "Epoch 15/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0591 - accuracy: 0.9798 - val_loss: 0.0590 - val_accuracy: 0.9468\n",
      "Epoch 16/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0555 - accuracy: 0.9828 - val_loss: 1.3359 - val_accuracy: 0.7421\n",
      "Epoch 17/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0480 - accuracy: 0.9844 - val_loss: 0.0426 - val_accuracy: 0.9749\n",
      "Epoch 18/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0573 - accuracy: 0.9815 - val_loss: 3.7292 - val_accuracy: 0.5566\n",
      "Epoch 19/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0481 - accuracy: 0.9842 - val_loss: 0.0267 - val_accuracy: 0.9588\n",
      "Epoch 20/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0551 - accuracy: 0.9824 - val_loss: 0.0363 - val_accuracy: 0.9461\n",
      "Epoch 21/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0478 - accuracy: 0.9835 - val_loss: 0.3391 - val_accuracy: 0.8200\n",
      "Epoch 22/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0503 - accuracy: 0.9838 - val_loss: 0.3068 - val_accuracy: 0.8036\n",
      "Epoch 23/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0460 - accuracy: 0.9841 - val_loss: 0.1791 - val_accuracy: 0.9377\n",
      "Epoch 24/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0416 - accuracy: 0.9857 - val_loss: 0.3893 - val_accuracy: 0.9337\n",
      "Epoch 25/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0371 - accuracy: 0.9878 - val_loss: 0.1491 - val_accuracy: 0.9683\n",
      "Epoch 26/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0381 - accuracy: 0.9879 - val_loss: 0.9026 - val_accuracy: 0.8270\n",
      "Epoch 27/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0439 - accuracy: 0.9849 - val_loss: 0.0948 - val_accuracy: 0.9719\n",
      "Epoch 28/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0377 - accuracy: 0.9879 - val_loss: 0.2767 - val_accuracy: 0.8601\n",
      "Epoch 29/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0397 - accuracy: 0.9882 - val_loss: 6.6217 - val_accuracy: 0.8798\n",
      "Epoch 30/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0378 - accuracy: 0.9880 - val_loss: 5.4076 - val_accuracy: 0.5366\n",
      "Epoch 31/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0472 - accuracy: 0.9842 - val_loss: 0.0326 - val_accuracy: 0.9913\n",
      "Epoch 32/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0339 - accuracy: 0.9896 - val_loss: 0.0134 - val_accuracy: 0.9909\n",
      "Epoch 33/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0252 - accuracy: 0.9918 - val_loss: 0.1769 - val_accuracy: 0.9741\n",
      "Epoch 34/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0288 - accuracy: 0.9905 - val_loss: 0.0278 - val_accuracy: 0.9654\n",
      "Epoch 35/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0328 - accuracy: 0.9887 - val_loss: 0.2154 - val_accuracy: 0.8474\n",
      "Epoch 36/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0328 - accuracy: 0.9894 - val_loss: 0.0271 - val_accuracy: 0.9814\n",
      "Epoch 37/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0272 - accuracy: 0.9911 - val_loss: 0.3508 - val_accuracy: 0.9690\n",
      "Epoch 38/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0286 - accuracy: 0.9908 - val_loss: 0.0042 - val_accuracy: 0.9880\n",
      "Epoch 39/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0266 - accuracy: 0.9911 - val_loss: 0.0151 - val_accuracy: 0.9858\n",
      "Epoch 40/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0305 - accuracy: 0.9898 - val_loss: 3.8962 - val_accuracy: 0.5144\n",
      "Epoch 41/50\n",
      "773/773 [==============================] - 8s 11ms/step - loss: 0.0267 - accuracy: 0.9924 - val_loss: 0.0802 - val_accuracy: 0.9832\n",
      "Epoch 42/50\n",
      "773/773 [==============================] - 8s 10ms/step - loss: 0.0232 - accuracy: 0.9927 - val_loss: 5.0927e-04 - val_accuracy: 0.9956\n",
      "\n",
      "Reached 99.5% accuracy so cancelling training!\n"
     ]
    }
   ],
   "source": [
    "history=model.fit(traindata_generator,epochs=50,validation_data=validationdata_generator,callbacks=[callback])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'r', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'r', label='Training Loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation Loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test accuracy: 99.81874227523804\n"
     ]
    }
   ],
   "source": [
    "print(\"test accuracy: \"+ str(model.evaluate_generator(testdata_generator)[1]*100))"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}