DIGIT_CLASSIFICATION_MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting tensorflow==2.12.0\n",
      "  Using cached tensorflow-2.12.0-cp311-cp311-win_amd64.whl.metadata (2.5 kB)\n",
      "Collecting tensorflow-gpu==2.12.0\n",
      "  Using cached tensorflow-gpu-2.12.0.tar.gz (2.6 kB)\n",
      "  Preparing metadata (setup.py): started\n",
      "  Preparing metadata (setup.py): finished with status 'done'\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Ignoring version 2.12.0 of tensorflow-gpu since it has invalid metadata:\n",
      "Requested tensorflow-gpu==2.12.0 from https://files.pythonhosted.org/packages/8a/45/fa31ced1db38f9424f262dfbf35747fe5378b5c808cecb373c8cb8e515d3/tensorflow-gpu-2.12.0.tar.gz has invalid metadata: Expected end or semicolon (after name and no valid version specifier)\n",
      "    python_version>\"3.7\"\n",
      "                  ^\n",
      "Please use pip<24.1 if you need to use this version.\n",
      "ERROR: Could not find a version that satisfies the requirement tensorflow-gpu==2.12.0 (from versions: 2.12.0)\n",
      "ERROR: No matching distribution found for tensorflow-gpu==2.12.0\n",
      "\n",
      "[notice] A new release of pip is available: 24.1.2 -> 24.2\n",
      "[notice] To update, run: C:\\Users\\Manya Aggarwal\\AppData\\Local\\Microsoft\\WindowsApps\\PythonSoftwareFoundation.Python.3.11_qbz5n2kfra8p0\\python.exe -m pip install --upgrade pip\n"
     ]
    }
   ],
   "source": [
    "!pip install tensorflow==2.12.0 tensorflow-gpu==2.12.0 opencv-python matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "import os\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "(x_train,y_train),(x_test,y_test)=tf.keras.datasets.mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "numpy.ndarray"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(x_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[[0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  ...\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]]\n",
      "\n",
      " [[0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  ...\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]]\n",
      "\n",
      " [[0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  ...\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]]\n",
      "\n",
      " ...\n",
      "\n",
      " [[0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  ...\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]]\n",
      "\n",
      " [[0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  ...\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]]\n",
      "\n",
      " [[0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  ...\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]\n",
      "  [0 0 0 ... 0 0 0]]]\n"
     ]
    }
   ],
   "source": [
    "print(x_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train=x_train.astype(\"float32\")/255.0\n",
    "x_test=x_test.astype(\"float32\")/255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(x_train[0],cmap=\"Blues\")\n",
    "plt.title(f'Label:{y_train[0]}')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[5 0 4 ... 5 6 8]\n"
     ]
    }
   ],
   "source": [
    "print(y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.99215686\n"
     ]
    }
   ],
   "source": [
    "print(x_train[0][7][13])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x600 with 16 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(6,6))\n",
    "for i in range(16):\n",
    "    plt.subplot(4,4,i+1)\n",
    "    plt.imshow(x_train[i],cmap=\"Blues\")\n",
    "    plt.title(f'Label:{y_train[i]}')\n",
    "    plt.axis('off')\n",
    "plt.tight_layout()    \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "#reshape the dataset for input in the dense layer of ANN\n",
    "x_train=x_train.reshape(x_train.shape[0],-1)\n",
    "x_test=x_test.reshape(x_test.shape[0],-1)\n",
    "\n",
    "#one hot encoding of labels \n",
    "y_train=tf.keras.utils.to_categorical(y_train,10)\n",
    "y_test=tf.keras.utils.to_categorical(y_test,10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 10)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Flatten"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_and_compile_model(activation_func,loss_func):\n",
    "    model= Sequential([\n",
    "        Flatten(input_shape=(784,)),\n",
    "        Dense(128,activation=activation_func),\n",
    "        Dense(64,activation=activation_func),\n",
    "        Dense(10,activation='softmax')\n",
    "    ])\n",
    "    model.compile(optimizer='adam',\n",
    "              loss=loss_func,\n",
    "               metrics=['accuracy'] )\n",
    "    return(model)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "activation function used:relu,Loss function used:categorical_crossentropy\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Manya Aggarwal\\AppData\\Local\\Packages\\PythonSoftwareFoundation.Python.3.11_qbz5n2kfra8p0\\LocalCache\\local-packages\\Python311\\site-packages\\keras\\src\\layers\\reshaping\\flatten.py:37: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n",
      "  super().__init__(**kwargs)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test_accuracy:0.9775\n",
      "activation function used:relu,Loss function used:mean_squared_error\n",
      "Test_accuracy:0.9772\n",
      "activation function used:tanh,Loss function used:categorical_crossentropy\n",
      "Test_accuracy:0.9772\n",
      "activation function used:tanh,Loss function used:mean_squared_error\n",
      "Test_accuracy:0.9775\n",
      "activation function used:sigmoid,Loss function used:categorical_crossentropy\n",
      "Test_accuracy:0.9774\n",
      "activation function used:sigmoid,Loss function used:mean_squared_error\n",
      "Test_accuracy:0.9756\n"
     ]
    }
   ],
   "source": [
    "activations=['relu','tanh','sigmoid']\n",
    "losses=['categorical_crossentropy','mean_squared_error']\n",
    "results=[]\n",
    "for activation in activations:\n",
    "    for loss in losses:\n",
    "        print(f\"activation function used:{activation},Loss function used:{loss}\")\n",
    "        model=build_and_compile_model(activation_func=activation,loss_func=loss)\n",
    "        history=model.fit(x_train,y_train,epochs=10,validation_data=(x_test,y_test),verbose=0)\n",
    "        _,accuracy=model.evaluate(x_test,y_test,verbose=0)\n",
    "        results.append({'Activation':activation,'Loss':loss,'Accuracy':accuracy})\n",
    "        print(f\"Test_accuracy:{accuracy:.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 10)"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Activation: relu, Loss: categorical_crossentropy, Accuracy: 0.9775\n",
      "Activation: relu, Loss: mean_squared_error, Accuracy: 0.9772\n",
      "Activation: tanh, Loss: categorical_crossentropy, Accuracy: 0.9772\n",
      "Activation: tanh, Loss: mean_squared_error, Accuracy: 0.9775\n",
      "Activation: sigmoid, Loss: categorical_crossentropy, Accuracy: 0.9774\n",
      "Activation: sigmoid, Loss: mean_squared_error, Accuracy: 0.9756\n"
     ]
    }
   ],
   "source": [
    "for result in results:\n",
    "    print(f\"Activation: {result['Activation']}, Loss: {result['Loss']}, Accuracy: {result['Accuracy']:.4f}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "same activation function and loss function but different layers and neurons in ANN\n",
    "\n",
    "activation function=Relu\n",
    "\n",
    "loss_function=categorical_cross_entropy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_and_compile_model(neuron_1,neuron_2):\n",
    "    model= Sequential([\n",
    "        Flatten(input_shape=(784,)),\n",
    "        Dense(neuron_1,activation='relu'),\n",
    "        Dense(neuron_2,activation='relu'),\n",
    "        Dense(10,activation='softmax')\n",
    "    ])\n",
    "    model.compile(optimizer='adam',\n",
    "              loss='categorical_crossentropy',\n",
    "               metrics=['accuracy'] )\n",
    "    return(model)   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "neuron in first layer:10,neuron in second layer:10\n",
      "Test_accuracy:0.9415\n",
      "neuron in first layer:30,neuron in second layer:30\n",
      "Test_accuracy:0.9636\n",
      "neuron in first layer:50,neuron in second layer:50\n",
      "Test_accuracy:0.9722\n",
      "neuron in first layer:75,neuron in second layer:75\n",
      "Test_accuracy:0.9735\n"
     ]
    }
   ],
   "source": [
    "neurons=[10,30,50,75]\n",
    "results=[]\n",
    "for neuron in neurons:\n",
    "        print(f\"neuron in first layer:{neuron},neuron in second layer:{neuron}\")\n",
    "        model=build_and_compile_model(neuron_1=neuron,neuron_2=neuron)\n",
    "        history=model.fit(x_train,y_train,epochs=10,validation_data=(x_test,y_test),verbose=0)\n",
    "        _,accuracy=model.evaluate(x_test,y_test,verbose=0)\n",
    "        results.append({'neurons_in_two_layers':neuron,'Accuracy':accuracy})\n",
    "        print(f\"Test_accuracy:{accuracy:.4f}\")"
   ]
  }
 ],
 "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}