Fashion MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import fashion_mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "(x_train,y_train),(x_test,y_test) = fashion_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape\n",
    "#has no color channel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x20713aea9e8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#grab single image\n",
    "single = x_train[10]\n",
    "plt.imshow(single, cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([9, 0, 0, ..., 3, 0, 5], dtype=uint8)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils.np_utils import to_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_cat_test = to_categorical(y_test,10) #no of class = 10\n",
    "y_cat_train = to_categorical(y_train,10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0., 0., 0., 0., 0., 0., 0., 0., 0., 1.], dtype=float32)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_cat_test[0] #one hot encoded"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "255"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "single.max() #still very high we have to normalize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#manually\n",
    "x_train = x_train/x_train.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#manually\n",
    "x_test = x_test/x_test.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test[0].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x2071d248c18>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(x_test[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((10000, 28, 28), (60000, 28, 28))"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test.shape, x_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#include color channel now it has only 1\n",
    "x_train = x_train.reshape(60000,28,28,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28, 1)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_test = x_test.reshape(10000,28,28,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 28, 28, 1)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Conv2D,MaxPool2D,Flatten "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\codehax41\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 25, 25, 32)        544       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 12, 12, 32)        0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 4608)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 128)               589952    \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 591,786\n",
      "Trainable params: 591,786\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "#CONV Layer\n",
    "model.add(Conv2D(filters=32,kernel_size=(4,4),input_shape=(28,28,1),activation='relu'))\n",
    "#Pooling Layer\n",
    "model.add(MaxPool2D(pool_size=(2,2)))\n",
    "#above 2 layers can understand 2d array\n",
    "#but dense layer only knows 1d so we need to flatten out\n",
    "model.add(Flatten())\n",
    "#dense\n",
    "model.add(Dense(128,activation='relu')) #no of neurons can choose any 128,64,256 etc\n",
    "model.add(Dense(10,activation='softmax')) #no of o/p\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "             optimizer='rmsprop',\n",
    "             metrics=['accuracy'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\codehax41\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\ops\\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\codehax41\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:1: UserWarning: The `nb_epoch` argument in `fit` has been renamed `epochs`.\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 48000 samples, validate on 12000 samples\n",
      "Epoch 1/2\n",
      "48000/48000 [==============================] - 64s 1ms/step - loss: 0.4105 - acc: 0.8542 - val_loss: 0.3183 - val_acc: 0.8875\n",
      "Epoch 2/2\n",
      "48000/48000 [==============================] - 58s 1ms/step - loss: 0.2828 - acc: 0.8990 - val_loss: 0.2950 - val_acc: 0.8942\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(x_train, y_cat_train, nb_epoch=2, validation_split=0.2, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 3s 275us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.31422339930534365, 0.8901]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(x_test,y_cat_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict_classes(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.80      0.83      1000\n",
      "           1       0.99      0.98      0.98      1000\n",
      "           2       0.80      0.86      0.83      1000\n",
      "           3       0.92      0.87      0.89      1000\n",
      "           4       0.74      0.92      0.82      1000\n",
      "           5       0.98      0.96      0.97      1000\n",
      "           6       0.75      0.62      0.68      1000\n",
      "           7       0.92      0.98      0.95      1000\n",
      "           8       0.97      0.98      0.98      1000\n",
      "           9       0.97      0.94      0.96      1000\n",
      "\n",
      "   micro avg       0.89      0.89      0.89     10000\n",
      "   macro avg       0.89      0.89      0.89     10000\n",
      "weighted avg       0.89      0.89      0.89     10000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test,predictions))"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}