Assignment04_part04.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "interracial-parish",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import cv2 as cv\n",
    "from tensorflow.keras.datasets import cifar10\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from tensorflow.keras.layers import Conv2D, MaxPooling2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "imperial-waterproof",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train: (50000, 32, 32, 3)\n",
      "x_test: (10000, 32, 32, 3)\n",
      "y_train: (50000, 10)\n",
      "y_test: (10000, 10)\n"
     ]
    }
   ],
   "source": [
    "(x_train, y_train), (x_test, y_test) = keras.datasets.cifar10.load_data()\n",
    "K = len(np.unique(y_train)) # Return the unique elements of a tratining output set and take it length as Classes\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "\n",
    "mean_image = np.mean(x_train, axis=(0,1,2,3))\n",
    "std = np.std(x_train, axis=(0,1,2,3))\n",
    "x_train = (x_train - mean_image)/(std+1e-7)\n",
    "x_test = (x_test - mean_image)/(std+1e-7)\n",
    "\n",
    "y_train = tf.keras.utils.to_categorical(y_train, num_classes=K)\n",
    "#This function returns a matrix of binary values (either ‘1’ or ‘0’). It has number of rows equal to the length of\n",
    "#the input vector and number of columns equal to the number of classes.\n",
    "y_test = tf.keras.utils.to_categorical(y_test, num_classes=K)\n",
    "\n",
    "print('x_train:', x_train.shape)\n",
    "print('x_test:', x_test.shape)\n",
    "print('y_train:', y_train.shape)\n",
    "print('y_test:', y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "incredible-syndicate",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_3 (Conv2D)            (None, 30, 30, 32)        896       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 15, 15, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 13, 13, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 6, 6, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 4, 4, 64)          36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 2, 2, 64)          0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 64)                16448     \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                650       \n",
      "=================================================================\n",
      "Total params: 73,418\n",
      "Trainable params: 73,418\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "# convolution layer & maxpooling layer with 32 nodes\n",
    "model.add(Conv2D(32, (3, 3), activation = \"relu\", input_shape=(32,32,3)))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "# convolution layer & maxpooling layer with 64 nodes\n",
    "model.add(Conv2D(64, (3, 3), activation = \"relu\"))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "model.add(Conv2D(64, (3, 3), activation = \"relu\"))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "# Flatten serves as a connection between the convolution and dense layers. \n",
    "model.add(Flatten()) \n",
    "# output layer\n",
    "model.add(Dense(64, activation = \"relu\"))\n",
    "model.add(Dense(10))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "addressed-madrid",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "1000/1000 [==============================] - 52s 51ms/step - loss: 1.6731 - accuracy: 0.3863 - val_loss: 1.2266 - val_accuracy: 0.5589\n",
      "Epoch 2/50\n",
      "1000/1000 [==============================] - 60s 60ms/step - loss: 1.1190 - accuracy: 0.6030 - val_loss: 0.9886 - val_accuracy: 0.6577\n",
      "Epoch 3/50\n",
      "1000/1000 [==============================] - 61s 61ms/step - loss: 0.9320 - accuracy: 0.6752 - val_loss: 0.9083 - val_accuracy: 0.6890\n",
      "Epoch 4/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.8158 - accuracy: 0.7144 - val_loss: 0.8581 - val_accuracy: 0.7048\n",
      "Epoch 5/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.7380 - accuracy: 0.7436 - val_loss: 0.8598 - val_accuracy: 0.7067\n",
      "Epoch 6/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.6811 - accuracy: 0.7616 - val_loss: 0.8284 - val_accuracy: 0.7173\n",
      "Epoch 7/50\n",
      "1000/1000 [==============================] - 44s 44ms/step - loss: 0.6141 - accuracy: 0.7835 - val_loss: 0.8292 - val_accuracy: 0.7179\n",
      "Epoch 8/50\n",
      "1000/1000 [==============================] - 44s 44ms/step - loss: 0.5740 - accuracy: 0.7963 - val_loss: 0.7885 - val_accuracy: 0.7343\n",
      "Epoch 9/50\n",
      "1000/1000 [==============================] - 42s 42ms/step - loss: 0.5300 - accuracy: 0.8143 - val_loss: 0.8301 - val_accuracy: 0.7284\n",
      "Epoch 10/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.5007 - accuracy: 0.8221 - val_loss: 0.8545 - val_accuracy: 0.7258\n",
      "Epoch 11/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.4731 - accuracy: 0.8342 - val_loss: 0.8196 - val_accuracy: 0.7350\n",
      "Epoch 12/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.4352 - accuracy: 0.8473 - val_loss: 0.8492 - val_accuracy: 0.7319\n",
      "Epoch 13/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.4058 - accuracy: 0.8547 - val_loss: 0.8921 - val_accuracy: 0.7240\n",
      "Epoch 14/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.3799 - accuracy: 0.8647 - val_loss: 0.9030 - val_accuracy: 0.7314\n",
      "Epoch 15/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.3504 - accuracy: 0.8757 - val_loss: 0.9382 - val_accuracy: 0.7238\n",
      "Epoch 16/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.3271 - accuracy: 0.8838 - val_loss: 0.9388 - val_accuracy: 0.7288\n",
      "Epoch 17/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.3090 - accuracy: 0.8886 - val_loss: 1.0084 - val_accuracy: 0.7234\n",
      "Epoch 18/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.2963 - accuracy: 0.8959 - val_loss: 1.0527 - val_accuracy: 0.7273\n",
      "Epoch 19/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.2773 - accuracy: 0.8996 - val_loss: 1.1067 - val_accuracy: 0.7212\n",
      "Epoch 20/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.2539 - accuracy: 0.9084 - val_loss: 1.0778 - val_accuracy: 0.7214\n",
      "Epoch 21/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.2450 - accuracy: 0.9111 - val_loss: 1.1805 - val_accuracy: 0.7159\n",
      "Epoch 22/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.2333 - accuracy: 0.9170 - val_loss: 1.2232 - val_accuracy: 0.7204\n",
      "Epoch 23/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.2142 - accuracy: 0.9210 - val_loss: 1.2353 - val_accuracy: 0.7231\n",
      "Epoch 24/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.2135 - accuracy: 0.9218 - val_loss: 1.3037 - val_accuracy: 0.7129\n",
      "Epoch 25/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1916 - accuracy: 0.9293 - val_loss: 1.3984 - val_accuracy: 0.7132\n",
      "Epoch 26/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1879 - accuracy: 0.9306 - val_loss: 1.4082 - val_accuracy: 0.7102\n",
      "Epoch 27/50\n",
      "1000/1000 [==============================] - 44s 44ms/step - loss: 0.1752 - accuracy: 0.9362 - val_loss: 1.4724 - val_accuracy: 0.7151\n",
      "Epoch 28/50\n",
      "1000/1000 [==============================] - 44s 44ms/step - loss: 0.1791 - accuracy: 0.9356 - val_loss: 1.4614 - val_accuracy: 0.7199\n",
      "Epoch 29/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1580 - accuracy: 0.9440 - val_loss: 1.4988 - val_accuracy: 0.7104\n",
      "Epoch 30/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1590 - accuracy: 0.9438 - val_loss: 1.6015 - val_accuracy: 0.7198\n",
      "Epoch 31/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1537 - accuracy: 0.9451 - val_loss: 1.6044 - val_accuracy: 0.7118\n",
      "Epoch 32/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1588 - accuracy: 0.9421 - val_loss: 1.5676 - val_accuracy: 0.7199\n",
      "Epoch 33/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1306 - accuracy: 0.9539 - val_loss: 1.6710 - val_accuracy: 0.7178\n",
      "Epoch 34/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1310 - accuracy: 0.9519 - val_loss: 1.7241 - val_accuracy: 0.7034\n",
      "Epoch 35/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.1331 - accuracy: 0.9528 - val_loss: 1.7150 - val_accuracy: 0.7182\n",
      "Epoch 36/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1326 - accuracy: 0.9519 - val_loss: 1.7710 - val_accuracy: 0.7094\n",
      "Epoch 37/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1240 - accuracy: 0.9542 - val_loss: 1.7890 - val_accuracy: 0.7195\n",
      "Epoch 38/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1258 - accuracy: 0.9553 - val_loss: 1.8700 - val_accuracy: 0.7120\n",
      "Epoch 39/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.1339 - accuracy: 0.9533 - val_loss: 1.8649 - val_accuracy: 0.7089\n",
      "Epoch 40/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1155 - accuracy: 0.9575 - val_loss: 1.8028 - val_accuracy: 0.7178\n",
      "Epoch 41/50\n",
      "1000/1000 [==============================] - 47s 47ms/step - loss: 0.1092 - accuracy: 0.9602 - val_loss: 1.9191 - val_accuracy: 0.7074\n",
      "Epoch 42/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1104 - accuracy: 0.9609 - val_loss: 2.0296 - val_accuracy: 0.7080\n",
      "Epoch 43/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1068 - accuracy: 0.9619 - val_loss: 2.0347 - val_accuracy: 0.7054\n",
      "Epoch 44/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.1054 - accuracy: 0.9629 - val_loss: 1.9833 - val_accuracy: 0.7182\n",
      "Epoch 45/50\n",
      "1000/1000 [==============================] - 46s 46ms/step - loss: 0.0993 - accuracy: 0.9649 - val_loss: 2.0484 - val_accuracy: 0.7137\n",
      "Epoch 46/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1121 - accuracy: 0.9622 - val_loss: 2.0992 - val_accuracy: 0.7030\n",
      "Epoch 47/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1000 - accuracy: 0.9653 - val_loss: 2.0756 - val_accuracy: 0.7130\n",
      "Epoch 48/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1054 - accuracy: 0.9642 - val_loss: 2.0693 - val_accuracy: 0.7114\n",
      "Epoch 49/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.0885 - accuracy: 0.9692 - val_loss: 2.1212 - val_accuracy: 0.7195\n",
      "Epoch 50/50\n",
      "1000/1000 [==============================] - 45s 45ms/step - loss: 0.1041 - accuracy: 0.9637 - val_loss: 2.1881 - val_accuracy: 0.7127\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='adam',loss=tf.keras.losses.CategoricalCrossentropy(from_logits=True),metrics=[\"accuracy\"])\n",
    "history=model.fit(x_train,y_train,epochs=50,batch_size=50,validation_data=(x_test,y_test),)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "spanish-intermediate",
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_history, test_loss_history, train_acc_history,val_acc_history = (history.history['loss'],history.history['val_loss'],history.history['accuracy'],history.history['val_accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "electric-crown",
   "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"
    }
   ],
   "source": [
    "fig, axes = plt.subplots()\n",
    "axes.plot(loss_history)  \n",
    "axes.set_xlabel(\"epochs\")\n",
    "axes.set_ylabel(\"LOSS\")\n",
    "plt.show()\n",
    "fig.savefig('LOSS.png',transparent=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "curious-halifax",
   "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"
    }
   ],
   "source": [
    "fig, axes = plt.subplots()\n",
    "axes.plot(test_loss_history)  \n",
    "axes.set_xlabel(\"epochs\")\n",
    "axes.set_ylabel(\"LOSS\")\n",
    "plt.show()\n",
    "fig.savefig('testing_loss.png',transparent=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "smoking-signature",
   "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"
    }
   ],
   "source": [
    "fig, axes = plt.subplots() \n",
    "axes.plot(train_acc_history)  \n",
    "axes.set_xlabel(\"epochs\")\n",
    "axes.set_ylabel(\"training accuracy\")\n",
    "plt.show()\n",
    "fig.savefig('training accuracy.png',transparent=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "careful-housing",
   "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"
    }
   ],
   "source": [
    "fig, axes = plt.subplots() \n",
    "axes.plot(val_acc_history)  \n",
    "axes.set_xlabel(\"epochs\")\n",
    "axes.set_ylabel(\"testing accuracy\")\n",
    "plt.show()\n",
    "fig.savefig('testing accuracy.png',transparent=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "behavioral-metropolitan",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}