HandwrittenDigitRecognition.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "64ff906f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "da8bc6db",
   "metadata": {},
   "outputs": [],
   "source": [
    "(x_train,y_train) , (x_test,y_test) = keras.datasets.mnist.load_data() # loading dataset "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "f6d82602",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "60000"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(x_train) # size of train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "bad3c110",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10000"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(x_test)  # size of test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "fd7c7f88",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(28, 28)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train[0].shape # shape of a single training example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "5a203637",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fc451097190>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(x_train[0]) # visualization of 1st training example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "9dfdb935",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[0] # what is the output of 1st traning example "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "308f00ec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5, 0, 4, 1, 9, 2], dtype=uint8)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[:6] # first 6 output of traning example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "d8171d27",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape # shape of whole training dataset "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "6630ebf5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# scaling data for better accuracy \n",
    "x_train = x_train / 255 \n",
    "x_test  = x_test / 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "d6a7cbbe",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train_flattened = x_train.reshape(len(x_train),28*28) # flattening the x_train to make it usable in DL model\n",
    "x_test_flattened = x_test.reshape(len(x_test),28*28)   # flattening the x_test to make it usable in DL model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "63e377a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 784)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test_flattened.shape # showing the modified shape of x_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "b00372d6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 784)"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train_flattened.shape # showing the modified shape of x_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "c95cacb7",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1875/1875 [==============================] - 4s 2ms/step - loss: 0.4721 - accuracy: 0.8765\n",
      "Epoch 2/5\n",
      "1875/1875 [==============================] - 3s 2ms/step - loss: 0.3041 - accuracy: 0.9158\n",
      "Epoch 3/5\n",
      "1875/1875 [==============================] - 3s 2ms/step - loss: 0.2835 - accuracy: 0.9205\n",
      "Epoch 4/5\n",
      "1875/1875 [==============================] - 3s 2ms/step - loss: 0.2731 - accuracy: 0.9241\n",
      "Epoch 5/5\n",
      "1875/1875 [==============================] - 3s 2ms/step - loss: 0.2667 - accuracy: 0.9254\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7fc451161a60>"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# creating a Deep Learning Model with Keras with only input layer and no hidden layer\n",
    "\n",
    "\n",
    "# Create a `Sequential` model and add a Dense layer as the first layer.\n",
    "\n",
    "model = keras.models.Sequential()\n",
    "model.add(keras.Input(shape = (784,)))\n",
    "model.add(keras.layers.Dense(10,activation = \"sigmoid\"))\n",
    "\n",
    "# Now the model will take as input arrays of shape (None, 784)\n",
    "# and output arrays of shape (None,10).\n",
    "\n",
    "#after creating a model we need to compile it with necessary arguments\n",
    "\n",
    "model.compile(\n",
    "    optimizer = 'adam',\n",
    "    loss = 'sparse_categorical_crossentropy',\n",
    "    metrics = ['accuracy']\n",
    ")\n",
    "\n",
    "model.fit(x_train_flattened,y_train,epochs=5)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "4d1c0354",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1875/1875 [==============================] - 5s 3ms/step - loss: 0.2770 - accuracy: 0.9222\n",
      "Epoch 2/5\n",
      "1875/1875 [==============================] - 4s 2ms/step - loss: 0.1262 - accuracy: 0.9621\n",
      "Epoch 3/5\n",
      "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0864 - accuracy: 0.9742\n",
      "Epoch 4/5\n",
      "1875/1875 [==============================] - 5s 2ms/step - loss: 0.0660 - accuracy: 0.9802\n",
      "Epoch 5/5\n",
      "1875/1875 [==============================] - 5s 2ms/step - loss: 0.0523 - accuracy: 0.9839\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7fc4487c0a90>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# let use add a hidden layer so that we can get e better accuracy \n",
    "\n",
    "\n",
    "# Create a `Sequential` model and add a Dense layer as the first layer and this time we add a hidden layer which has (None,10) shape.\n",
    "\n",
    "model = keras.models.Sequential()\n",
    "# model.add(keras.Input(shape = (784,))) if we can flattened our train dataset earlier\n",
    "model.add(keras.Input(shape = (28,28)))\n",
    "model.add(keras.layers.Flatten()) # if we didn't flattern our train dataset earlier then using this method we can flatten it\n",
    "model.add(keras.layers.Dense(100,activation = \"relu\"))\n",
    "model.add(keras.layers.Dense(10,activation = \"sigmoid\"))\n",
    "\n",
    "\n",
    "# Now the model will take as input arrays of shape (None, 784)\n",
    "# and output arrays of shape (None,10).\n",
    "\n",
    "#after creating a model we need to compile it with necessary arguments\n",
    "\n",
    "model.compile(\n",
    "    optimizer = 'adam',\n",
    "    loss = 'sparse_categorical_crossentropy',\n",
    "    metrics = ['accuracy']\n",
    ")\n",
    "\n",
    "# model.fit(x_train_flattened,y_train,epochs=5) # if we have flattend earlier \n",
    "model.fit(x_train,y_train,epochs=5)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "a2140a1b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 4ms/step - loss: 0.0748 - accuracy: 0.9778\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.07481581717729568, 0.9778000116348267]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(x_test,y_test) # if you are using the Flatten() function provided by keras\n",
    "\n",
    "# if you are flattening the train dataset earlier so your model is not using the flatten() method\n",
    "# model.evaluate(x_test_flattened,y_test) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "9b4f513c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 2ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([3.4183193e-02, 3.7986778e-03, 3.4818786e-01, 8.8238591e-01,\n",
       "       8.4784034e-07, 1.5125720e-02, 3.6207351e-10, 9.9997956e-01,\n",
       "       5.2689321e-02, 2.2721575e-01], dtype=float32)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_prediction = model.predict(x_test) # predicting the whole test data\n",
    "y_prediction[0] # showing the 0th prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "e199a93f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(y_prediction[0]) # max possibility among 10 digits"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "d5cec1d1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(x_test[0]); # showing the 0th test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "3f5bc962",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[7, 2, 1, 0, 4]"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# for every prediction we need to know the highest similarity score because prediction is providing us with a array of size 10\n",
    "y_prediction_labels = [np.argmax(i) for i in y_prediction]\n",
    "y_prediction_labels[:5] # displaying first 5 prediction with highest similarity digit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "39e9fb75",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(10, 10), dtype=int32, numpy=\n",
       "array([[ 968,    0,    0,    1,    0,    0,    3,    1,    3,    4],\n",
       "       [   0, 1125,    4,    0,    0,    0,    2,    1,    3,    0],\n",
       "       [   3,    1, 1014,    0,    2,    0,    2,    6,    4,    0],\n",
       "       [   1,    0,   10,  976,    1,    2,    0,    7,    7,    6],\n",
       "       [   2,    0,    3,    1,  957,    0,    3,    3,    2,   11],\n",
       "       [   4,    1,    1,    7,    2,  859,    7,    2,    5,    4],\n",
       "       [   5,    3,    4,    0,    2,    4,  938,    0,    2,    0],\n",
       "       [   2,    3,    6,    0,    0,    0,    0, 1014,    1,    2],\n",
       "       [   8,    0,    3,    5,    6,    2,    2,    3,  943,    2],\n",
       "       [   1,    3,    0,    3,    9,    1,    1,    6,    1,  984]],\n",
       "      dtype=int32)>"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# confusion matrix\n",
    "\n",
    "cm = tf.math.confusion_matrix(labels = y_test, predictions = y_prediction_labels )\n",
    "cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "0ea23a1c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(69.0, 0.5, 'Truth')"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualizing the confusion matrix in a appealing way for this we need to import seaborn\n",
    "\n",
    "import seaborn as sn\n",
    "plt.figure(figsize = (10,7))\n",
    "sn.heatmap(cm,annot = True , fmt = 'd')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "52c570d7",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  },
  "vscode": {
   "interpreter": {
    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}