CNN_fashioMNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Using Convolutional Neural Network for Classification problem. \n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from keras.datasets import fashion_mnist\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "from sklearn import preprocessing\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "This dataset has 10 classes\n"
     ]
    }
   ],
   "source": [
    "# dataset's information https://www.tensorflow.org/api_docs/python/tf/keras/datasets/fashion_mnist/load_data\n",
    "(x_train, y_train), (x_test, y_test) = fashion_mnist.load_data()\n",
    "\n",
    "# Finding the number of classes\n",
    "classes = np.unique(y_train)\n",
    "nClasses = len(classes)\n",
    "print(\"This dataset has {0:d} classes\".format(nClasses))\n",
    "\n",
    "# Defining a dictionaty for label name\n",
    "y_names = {0: \"T-shirt\", 1: \"Trouser\", 2: \"Pullover\", 3: \"Dress\", 4: \"Coat\", 5: \"Sandal\", 6: \"Shirt\", 7: \"Sneaker\", 8: \"Bag\", 9: \"Ankle Boot\"}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The dimention of each image is 28x28, the training dataset has 60000 sample and testing dataset has 10000 samples\n"
     ]
    }
   ],
   "source": [
    "n_traing_samples = len(x_train)\n",
    "n_test_sample = len(x_test)\n",
    "dimension_image = x_train[0].shape \n",
    "print(\"The dimention of each image is {0:d}x{1:d}, the training dataset has {2:d} sample and testing dataset has {3:d} samples\".format(dimension_image[0],dimension_image[1],n_traing_samples,n_test_sample))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Showing the first 3 images\n",
    "_, fig = plt.subplots(nrows=1, ncols=3, figsize=(10, 10))\n",
    "for ax, image, label in zip(fig, x_train, y_train):\n",
    "    ax.set_axis_off()\n",
    "    ax.imshow(image, cmap=plt.cm.gray_r)\n",
    "    ax.set_title('Training: %s' % y_names[label])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# To use a CNN model the input's format must be (width x high x channels),\n",
    "train_X = x_train.reshape((-1, 28,28, 1))\n",
    "test_X = x_test.reshape((-1, 28,28, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Changing format from int8 to float32 (Best numerical precision floatingPoints) and normalizing the data.\n",
    "train_X = train_X.astype('float32')\n",
    "test_X = test_X.astype('float32')\n",
    "train_X = train_X / 255\n",
    "test_X = test_X / 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Transform your labels values into a categorial format or best known as one_hot encoding.\n",
    "train_Y_one_hot = to_categorical(y_train)\n",
    "test_Y_one_hot = to_categorical(y_test)\n",
    "# use train_test_split function to crear validation training set from train data.\n",
    "# Split data into 80% train and 20% test subsets\n",
    "X_train_model, X_validation, y_train_model, y_validation = train_test_split(train_X, train_Y_one_hot, test_size=0.2, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing the required layers operation and activation function to build your model\n",
    "from keras.models import Sequential,Input,Model\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.layers.advanced_activations import LeakyReLU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining basic parameters\n",
    "batch_size = 32\n",
    "epochs = 15\n",
    "num_classes = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Buidling your CNN model\n",
    "fashion_model = Sequential()\n",
    "fashion_model.add(Conv2D(32, kernel_size=(3, 3),activation='linear',input_shape=(28,28,1),padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))\n",
    "fashion_model.add(MaxPooling2D((2, 2),padding='same'))\n",
    "fashion_model.add(Conv2D(64, (3, 3), activation='linear',padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))\n",
    "fashion_model.add(MaxPooling2D(pool_size=(2, 2),padding='same'))\n",
    "fashion_model.add(Conv2D(128, (3, 3), activation='linear',padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))                  \n",
    "fashion_model.add(MaxPooling2D(pool_size=(2, 2),padding='same'))\n",
    "fashion_model.add(Flatten())\n",
    "fashion_model.add(Dense(128, activation='linear'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))                  \n",
    "fashion_model.add(Dense(num_classes, activation='softmax')) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compiling your model\n",
    "fashion_model.compile(loss='binary_crossentropy', optimizer='adam',metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 28, 28, 32)        320       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu (LeakyReLU)      (None, 28, 28, 32)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 14, 14, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 14, 14, 64)        18496     \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_1 (LeakyReLU)    (None, 14, 14, 64)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 7, 7, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 7, 7, 128)         73856     \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_2 (LeakyReLU)    (None, 7, 7, 128)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 4, 4, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 128)               262272    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_3 (LeakyReLU)    (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 356,234\n",
      "Trainable params: 356,234\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Verify the outpur shape for each layer and the number of trainable parameters.\n",
    "fashion_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "1500/1500 [==============================] - 72s 37ms/step - loss: 0.1323 - accuracy: 0.7364 - val_loss: 0.0593 - val_accuracy: 0.8885\n",
      "Epoch 2/15\n",
      "1500/1500 [==============================] - 51s 34ms/step - loss: 0.0540 - accuracy: 0.8951 - val_loss: 0.0483 - val_accuracy: 0.9056\n",
      "Epoch 3/15\n",
      "1500/1500 [==============================] - 51s 34ms/step - loss: 0.0439 - accuracy: 0.9148 - val_loss: 0.0439 - val_accuracy: 0.9142\n",
      "Epoch 4/15\n",
      "1500/1500 [==============================] - 49s 33ms/step - loss: 0.0375 - accuracy: 0.9275 - val_loss: 0.0438 - val_accuracy: 0.9148\n",
      "Epoch 5/15\n",
      "1500/1500 [==============================] - 49s 33ms/step - loss: 0.0321 - accuracy: 0.9376 - val_loss: 0.0416 - val_accuracy: 0.9193\n",
      "Epoch 6/15\n",
      "1500/1500 [==============================] - 48s 32ms/step - loss: 0.0278 - accuracy: 0.9466 - val_loss: 0.0407 - val_accuracy: 0.9208\n",
      "Epoch 7/15\n",
      "1500/1500 [==============================] - 49s 33ms/step - loss: 0.0245 - accuracy: 0.9538 - val_loss: 0.0421 - val_accuracy: 0.9205\n",
      "Epoch 8/15\n",
      "1500/1500 [==============================] - 49s 32ms/step - loss: 0.0212 - accuracy: 0.9612 - val_loss: 0.0463 - val_accuracy: 0.9201\n",
      "Epoch 9/15\n",
      "1500/1500 [==============================] - 54s 36ms/step - loss: 0.0170 - accuracy: 0.9697 - val_loss: 0.0468 - val_accuracy: 0.9206\n",
      "Epoch 10/15\n",
      "1500/1500 [==============================] - 55s 37ms/step - loss: 0.0148 - accuracy: 0.9737 - val_loss: 0.0502 - val_accuracy: 0.9219\n",
      "Epoch 11/15\n",
      "1500/1500 [==============================] - 60s 40ms/step - loss: 0.0125 - accuracy: 0.9790 - val_loss: 0.0590 - val_accuracy: 0.9151\n",
      "Epoch 12/15\n",
      "1500/1500 [==============================] - 57s 38ms/step - loss: 0.0112 - accuracy: 0.9804 - val_loss: 0.0550 - val_accuracy: 0.9246\n",
      "Epoch 13/15\n",
      "1500/1500 [==============================] - 48s 32ms/step - loss: 0.0092 - accuracy: 0.9847 - val_loss: 0.0631 - val_accuracy: 0.9216\n",
      "Epoch 14/15\n",
      "1500/1500 [==============================] - 48s 32ms/step - loss: 0.0083 - accuracy: 0.9866 - val_loss: 0.0670 - val_accuracy: 0.9233\n",
      "Epoch 15/15\n",
      "1500/1500 [==============================] - 48s 32ms/step - loss: 0.0073 - accuracy: 0.9882 - val_loss: 0.0672 - val_accuracy: 0.9237\n"
     ]
    }
   ],
   "source": [
    "# Train your model\r\n",
    "fashion_trained_history = fashion_model.fit(X_train_model, y_train_model, batch_size=batch_size, epochs=epochs, verbose=1, validation_data=(X_validation, y_validation))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 3s 8ms/step - loss: 0.0716 - accuracy: 0.9185\n",
      "The loss of the model is 0.071647, and the accurracy is 0.918500\n"
     ]
    }
   ],
   "source": [
    "# Evaluating the performance of the model usign the validation data\n",
    "test_eval = fashion_model.evaluate(test_X, test_Y_one_hot, verbose=1)\n",
    "print(\"The loss of the model is {0:f}, and the accurracy is {1:f}\".format(test_eval[0],test_eval[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# using fashion_trained_history information to provide the performance of the model during the training stage\n",
    "accuracy = fashion_trained_history.history['accuracy']\n",
    "val_accuracy = fashion_trained_history.history['val_accuracy']\n",
    "loss = fashion_trained_history.history['loss']\n",
    "val_loss = fashion_trained_history.history['val_loss']\n",
    "epochs = range(len(accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(epochs, accuracy, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_accuracy, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000,)\n",
      "(10000,)\n"
     ]
    }
   ],
   "source": [
    "# Predicting the new data \"test_X, y_test\"\n",
    "predicted = fashion_model.predict(test_X)\n",
    "\n",
    "# taking the max argument\n",
    "predicted = np.argmax(np.round(predicted),axis=1)\n",
    "print(predicted.shape)\n",
    "print(y_test.shape)\n",
    "#print(\"size fo predicted array {0:d}x{1:d} and size of y_test array {2:d}\".format(predicted.shape[0],predicted.shape[1],y_test.shape))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 9179 correct labels\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Now let's select those samples that were correctly classified.\n",
    "correct = np.where(predicted==y_test)[0]\n",
    "print(\"Found %d correct labels\" % len(correct))\n",
    "for i, correct in enumerate(correct[:9]):\n",
    "    plt.subplot(3,3,i+1)\n",
    "    plt.imshow(test_X[correct].reshape(28,28), cmap='gray', interpolation='none')\n",
    "    plt.title(\"Predicted {}, Class {}\".format(predicted[correct], y_test[correct]))\n",
    "    plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 821 incorrect labels\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Now let's select those samples that were incorrectly classified.\n",
    "incorrect = np.where(predicted!=y_test)[0]\n",
    "print(\"Found %d incorrect labels\" % len(incorrect))\n",
    "for i, incorrect in enumerate(incorrect[:9]):\n",
    "    plt.subplot(3,3,i+1)\n",
    "    plt.imshow(test_X[incorrect].reshape(28,28), cmap='gray', interpolation='none')\n",
    "    plt.title(\"Predicted {}, Class {}\".format(predicted[incorrect], y_test[incorrect]))\n",
    "    plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "     Class 0       0.84      0.86      0.85      1000\n",
      "     Class 1       0.99      0.98      0.99      1000\n",
      "     Class 2       0.87      0.88      0.88      1000\n",
      "     Class 3       0.91      0.93      0.92      1000\n",
      "     Class 4       0.89      0.84      0.86      1000\n",
      "     Class 5       1.00      0.97      0.98      1000\n",
      "     Class 6       0.76      0.77      0.77      1000\n",
      "     Class 7       0.95      0.98      0.97      1000\n",
      "     Class 8       0.98      0.98      0.98      1000\n",
      "     Class 9       0.97      0.97      0.97      1000\n",
      "\n",
      "    accuracy                           0.92     10000\n",
      "   macro avg       0.92      0.92      0.92     10000\n",
      "weighted avg       0.92      0.92      0.92     10000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#Printing Classification reports\n",
    "from sklearn.metrics import classification_report\n",
    "target_names = [\"Class {}\".format(i) for i in range(num_classes)]\n",
    "print(classification_report(y_test, predicted, target_names=target_names))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "346362102b7552e91385457e3bb81cf6a30a33df4987af1bb84d68a48aef1f8e"
  },
  "kernelspec": {
   "display_name": "Python 3.9.5 64-bit (windows store)",
   "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": "3.9.5"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}