Neural_Network_from_Scratch.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "112169eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "618289d2",
   "metadata": {},
   "source": [
    "## Part 1 : Get the data and Preprocess it "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "de613434",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 1. Load the data from the .csv file, into a pandas dataframe (pandas makes loading csv files easy)\n",
    "pd_data = pd.read_csv('./mnist_data/train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "f1cb35d6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>label</th>\n",
       "      <th>pixel0</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel774</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 785 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   label  pixel0  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  \\\n",
       "0      1       0       0       0       0       0       0       0       0   \n",
       "1      0       0       0       0       0       0       0       0       0   \n",
       "2      1       0       0       0       0       0       0       0       0   \n",
       "3      4       0       0       0       0       0       0       0       0   \n",
       "4      0       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel8  ...  pixel774  pixel775  pixel776  pixel777  pixel778  pixel779  \\\n",
       "0       0  ...         0         0         0         0         0         0   \n",
       "1       0  ...         0         0         0         0         0         0   \n",
       "2       0  ...         0         0         0         0         0         0   \n",
       "3       0  ...         0         0         0         0         0         0   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel780  pixel781  pixel782  pixel783  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2         0         0         0         0  \n",
       "3         0         0         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 785 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "fd4e8252",
   "metadata": {},
   "outputs": [],
   "source": [
    "# a - convert the pandas dataframe to a numpy dataframe\n",
    "np_data = np.array(pd_data)\n",
    "\n",
    "\n",
    "# b - get the total number of rows and columns, and shuffle the dataset\n",
    "rows, cols = np_data.shape\n",
    "np.random.shuffle(np_data)\n",
    "\n",
    "\n",
    "# c - now we only want to work with 1000 test and 41,000 training examples\n",
    "test_data = np_data[0:1000].T # this get the first 1000 rows, and all their colums and transposes it \n",
    "    #  -- the end result is a  matric of (785-rows x 1000-cols), \n",
    "    #  --- where the rows = pixels + class, and cols = images (i.e. each column is one image)\n",
    "    \n",
    "    \n",
    "# d - next we seperate the target values from the pixels \n",
    "y_test = test_data[0]\n",
    "x_test = test_data[1: ]\n",
    "x_test = x_test / 255.0 \n",
    "\n",
    "''' WARNING !!!!! If you dont normalize, training will fail because of nan values -- \n",
    "        i.e when you update the values during backpropigation, since the values are large this\n",
    "            causes some values to be so large they become nan's\n",
    "'''\n",
    "\n",
    "# e - and we do the same for the training data\n",
    "train_data = np_data[1000:].T\n",
    "y_train = train_data[0]\n",
    "x_train = train_data[1:]\n",
    "x_train = x_train / 255.0\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ed482c1d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1. Train Data shape : (784, 41000)\n",
      "2. Test Data shape : (784, 1000)\n"
     ]
    }
   ],
   "source": [
    "# f - get the shapes of the test and train data\n",
    "print(f\"1. Train Data shape : {x_train.shape}\")\n",
    "print(f\"2. Test Data shape : {x_test.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "3a40af7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# g - we need to one hot encode the Y / Target values\n",
    "def one_hot_encode(y):\n",
    "    ''' Notes: \n",
    "        1 - this creates a one-hot colum vector for each value in the column vector, creating\n",
    "            a matrix where each column is a one hot encoding of some number(digit) the picture is representing\n",
    "            \n",
    "        2 - this is a fast method of computing one hot encoding by using :\n",
    "              a - np.arange - creates an array from 0 to the y.size(number of examples)\n",
    "              b - [np.arange(y.size), y] - tells us the row we want to acess (np.arrange), \n",
    "                   and its correspoding column - which we want to set as 1 since its the proper class\n",
    "    '''\n",
    "    one_hot_y = np.zeros( (y.size, y.max()+1) ) # y.size = m (number of examples)\n",
    "                                                # y.max() - is 9 because we assume 9 is the highest number\n",
    "    \n",
    "    one_hot_y[np.arange(y.size), y] = 1 \n",
    "    one_hot_y = one_hot_y.T \n",
    "    return one_hot_y # we want each col to be an example rather than each row\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c5e254a1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "38d14be6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.25       0.04444444]\n",
      " [0.75       0.95555556]]\n",
      "[[0.02040816 0.04081633]\n",
      " [0.06122449 0.87755102]]\n"
     ]
    }
   ],
   "source": [
    "# side note == see the difference between the sum() - sums the cols -  vs np.sum() function - which sums all the values in the np.array\n",
    "a = np.array(( [1,2],\n",
    "               [3,43]  ))\n",
    "print(a/ sum(a))\n",
    "print(a/ np.sum(a))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2f0b83d2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "27fb7d4a",
   "metadata": {},
   "source": [
    "## Part 2 - create the neural network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "f8204051",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\"\"\" Step 1 - Initalize all the parameters \"\"\"\n",
    "def initalize_params():\n",
    "    # layer 1\n",
    "    W_1 = np.random.rand(10,784) - 0.5 \n",
    "    b_1 = np.random.rand(10,1) - 0.5 \n",
    "    \n",
    "    # layer 2\n",
    "    W_2 = np.random.rand(10,10) - 0.5 \n",
    "    b_2 = np.random.rand(10,1) - 0.5 \n",
    "    \n",
    "    '''  Notes :\n",
    "            use rand - (dist. between 0 and 1) not randn - (dist between -0.5 and 0.5) \n",
    "            and we subtract 0.5 from them because we want the values to be between -0.5 and 0.5\n",
    "             qq -- why??\n",
    "    '''\n",
    "    return W_1, b_1, W_2, b_2\n",
    "\n",
    "\n",
    "\n",
    "\"\"\" Step 2 - Forward propagation \"\"\"\n",
    "def ReLU(Z):\n",
    "    return np.maximum(Z, 0) # this is element wise application of ReLU\n",
    "    \n",
    "def SoftMax(Z):\n",
    "    A = np.exp(Z)/sum(np.exp(Z)) \n",
    "    return A\n",
    "    # np.exp - is done element-wise\n",
    "    # sum(np.array(..)) - keeps the same # of cols and colapes the rows --  which is what we want\n",
    "    # use sum not np.sum in order that the sum will be taken over a single column, not the whole matrix\n",
    "    \n",
    "def forward_propigation(W_1, b_1, W_2, b_2, X):\n",
    "    \n",
    "    # 1 - Input -> First Layer of Nodes\n",
    "    Z_1 = W_1.dot(X) + b_1\n",
    "    A_1 = ReLU(Z_1)\n",
    "    \n",
    "    # 2 - First Layer of Nodes -> Second Layer of Nodes    \n",
    "    Z_2 = W_2.dot(A_1) + b_2\n",
    "    A_2 = SoftMax(Z_2)\n",
    "    \n",
    "    return Z_1, A_1, Z_2 ,A_2\n",
    "\n",
    "\n",
    "\"\"\" Step 3 - Backwards propagation \"\"\"\n",
    "def derivative_ReLU(Z):\n",
    "    # for x > 0 its one, otherwise its zero - \n",
    "    #  - the z>0 is done element wise and true=1 and false=0, so this is a fast method of calcualting the derivative or ReLU\n",
    "    return Z > 0\n",
    "\n",
    "def backward_propigation(Z_1, A_1, Z_2 ,A_2, W_2, X, Y):\n",
    "    one_hot_of_y = one_hot_encode(Y)\n",
    "    m = Y.size\n",
    "\n",
    "    dZ_2 = A_2 - one_hot_of_y\n",
    "\n",
    "    dW_2 = ( 1/m ) * dZ_2.dot(A_1.T)\n",
    "    \n",
    "    db_2 = ( 1/m ) * np.sum(dZ_2)\n",
    "    \n",
    "    dZ_1 = W_2.T.dot(dZ_2) * derivative_ReLU(Z_1)\n",
    "    \n",
    "    dW_1 = ( 1/m ) * dZ_1.dot(X.T)\n",
    "    \n",
    "    db_1 = ( 1/m ) * np.sum(dZ_1)\n",
    "    \n",
    "    return dW_1, db_1, dW_2, db_2\n",
    "\n",
    "\n",
    "\"\"\" Step 4 - Update the Parameters \"\"\"\n",
    "def update_params(W_1, b_1, W_2, b_2, dW_1, db_1, dW_2, db_2, alpha):\n",
    "    W_1 = W_1 - alpha * dW_1\n",
    "    b_1 = b_1 - alpha * db_1\n",
    "    W_2 = W_2 - alpha * dW_2\n",
    "    b_2 = b_2 - alpha * db_2\n",
    "    \n",
    "    return W_1, b_1, W_2, b_2\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3b6b174b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "da995205",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\" The functions to compute  (1) predictions, (2) accuracy and (3) loss \"\"\"\n",
    "def get_predictions(A_2):\n",
    "    return np.argmax(A_2, axis=0) # axis=0 - for the max value in the columns\n",
    "    # since the value with max probability will be the class that we predict\n",
    "    \n",
    "def get_accuracy(predictions, ground_truth):\n",
    "    #print(predictions, ground_truth)\n",
    "    return np.sum(predictions==ground_truth) / ground_truth.size\n",
    "\n",
    "def get_loss(predicted, ground_truth):\n",
    "    onehot_y = one_hot_encode(ground_truth)\n",
    "    return abs(np.sum(onehot_y-predicted))\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "951d3702",
   "metadata": {},
   "outputs": [],
   "source": [
    "def gradiant_decent(X, Y, num_iterations, alpha):\n",
    "    # step 1 - initalize our parameters\n",
    "    W_1, b_1, W_2, b_2 = initalize_params()\n",
    "    loss_history = []\n",
    "    accuracy_history = []\n",
    "    \n",
    "    # step 2 - run gradiant decent on the our neural network for num_itterations\n",
    "    for i in range(num_iterations):\n",
    "        \n",
    "        # a - forward propigation\n",
    "        Z_1, A_1, Z_2 ,A_2 = forward_propigation(W_1, b_1, W_2, b_2, X)\n",
    "        \n",
    "        # b - backward propigation \n",
    "        dW_1, db_1, dW_2, db_2 = backward_propigation(Z_1, A_1, Z_2 ,A_2, W_2, X, Y)\n",
    "        \n",
    "        \n",
    "        # c - update parameters\n",
    "        W_1, b_1, W_2, b_2 = update_params(W_1, b_1, W_2, b_2, dW_1, db_1, dW_2, db_2, alpha)\n",
    "        \n",
    "        \n",
    "        # d - print out the accuracy - every 50 itterations\n",
    "        if i%10 == 0:\n",
    "            print(f\"Iteration  #{i}\")\n",
    "            print(f\"-- Accuracy score : {get_accuracy(get_predictions(A_2), Y)}\")\n",
    "\n",
    "        # e - calculate the loss and accuracy of the model for the current itteration\n",
    "        loss_history.append(get_loss(A_2, Y)) \n",
    "        accuracy_history.append(get_accuracy(get_predictions(A_2), Y))\n",
    "        \n",
    "    return W_1, b_1, W_2, b_2, loss_history, accuracy_history\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab0e85c6",
   "metadata": {},
   "source": [
    "## Part 3 - Train the Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "9b4fe37f",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration  #0\n",
      "-- Accuracy score : 0.09195121951219512\n",
      "Iteration  #10\n",
      "-- Accuracy score : 0.20882926829268292\n",
      "Iteration  #20\n",
      "-- Accuracy score : 0.29892682926829267\n",
      "Iteration  #30\n",
      "-- Accuracy score : 0.37573170731707317\n",
      "Iteration  #40\n",
      "-- Accuracy score : 0.4427317073170732\n",
      "Iteration  #50\n",
      "-- Accuracy score : 0.492390243902439\n",
      "Iteration  #60\n",
      "-- Accuracy score : 0.5394634146341464\n",
      "Iteration  #70\n",
      "-- Accuracy score : 0.5805365853658536\n",
      "Iteration  #80\n",
      "-- Accuracy score : 0.6150731707317073\n",
      "Iteration  #90\n",
      "-- Accuracy score : 0.643\n",
      "Iteration  #100\n",
      "-- Accuracy score : 0.666609756097561\n",
      "Iteration  #110\n",
      "-- Accuracy score : 0.6875853658536585\n",
      "Iteration  #120\n",
      "-- Accuracy score : 0.7044634146341463\n",
      "Iteration  #130\n",
      "-- Accuracy score : 0.7180243902439024\n",
      "Iteration  #140\n",
      "-- Accuracy score : 0.7309268292682927\n",
      "Iteration  #150\n",
      "-- Accuracy score : 0.7411219512195122\n",
      "Iteration  #160\n",
      "-- Accuracy score : 0.7508048780487805\n",
      "Iteration  #170\n",
      "-- Accuracy score : 0.7599024390243903\n",
      "Iteration  #180\n",
      "-- Accuracy score : 0.7676341463414634\n",
      "Iteration  #190\n",
      "-- Accuracy score : 0.7742439024390244\n",
      "Iteration  #200\n",
      "-- Accuracy score : 0.7806585365853659\n",
      "Iteration  #210\n",
      "-- Accuracy score : 0.786609756097561\n",
      "Iteration  #220\n",
      "-- Accuracy score : 0.7917804878048781\n",
      "Iteration  #230\n",
      "-- Accuracy score : 0.7962439024390244\n",
      "Iteration  #240\n",
      "-- Accuracy score : 0.800829268292683\n",
      "Iteration  #250\n",
      "-- Accuracy score : 0.8046829268292683\n",
      "Iteration  #260\n",
      "-- Accuracy score : 0.8083170731707318\n",
      "Iteration  #270\n",
      "-- Accuracy score : 0.8124146341463415\n",
      "Iteration  #280\n",
      "-- Accuracy score : 0.8161951219512195\n",
      "Iteration  #290\n",
      "-- Accuracy score : 0.8194634146341463\n",
      "Iteration  #300\n",
      "-- Accuracy score : 0.8226829268292682\n",
      "Iteration  #310\n",
      "-- Accuracy score : 0.8255853658536585\n",
      "Iteration  #320\n",
      "-- Accuracy score : 0.827780487804878\n",
      "Iteration  #330\n",
      "-- Accuracy score : 0.8302682926829268\n",
      "Iteration  #340\n",
      "-- Accuracy score : 0.8327317073170731\n",
      "Iteration  #350\n",
      "-- Accuracy score : 0.8349756097560975\n",
      "Iteration  #360\n",
      "-- Accuracy score : 0.8365853658536585\n",
      "Iteration  #370\n",
      "-- Accuracy score : 0.8388536585365853\n",
      "Iteration  #380\n",
      "-- Accuracy score : 0.8409268292682927\n",
      "Iteration  #390\n",
      "-- Accuracy score : 0.8428536585365853\n",
      "Iteration  #400\n",
      "-- Accuracy score : 0.8442682926829268\n",
      "Iteration  #410\n",
      "-- Accuracy score : 0.845780487804878\n",
      "Iteration  #420\n",
      "-- Accuracy score : 0.8476829268292683\n",
      "Iteration  #430\n",
      "-- Accuracy score : 0.8492682926829268\n",
      "Iteration  #440\n",
      "-- Accuracy score : 0.8504390243902439\n",
      "Iteration  #450\n",
      "-- Accuracy score : 0.8520975609756097\n",
      "Iteration  #460\n",
      "-- Accuracy score : 0.8536829268292683\n",
      "Iteration  #470\n",
      "-- Accuracy score : 0.8550975609756097\n",
      "Iteration  #480\n",
      "-- Accuracy score : 0.8564634146341463\n",
      "Iteration  #490\n",
      "-- Accuracy score : 0.8577560975609756\n",
      "Iteration  #500\n",
      "-- Accuracy score : 0.8587560975609756\n",
      "Iteration  #510\n",
      "-- Accuracy score : 0.8598536585365854\n",
      "Iteration  #520\n",
      "-- Accuracy score : 0.861\n",
      "Iteration  #530\n",
      "-- Accuracy score : 0.8622195121951219\n",
      "Iteration  #540\n",
      "-- Accuracy score : 0.8632682926829268\n",
      "Iteration  #550\n",
      "-- Accuracy score : 0.8642195121951219\n",
      "Iteration  #560\n",
      "-- Accuracy score : 0.8653414634146341\n",
      "Iteration  #570\n",
      "-- Accuracy score : 0.8659512195121951\n",
      "Iteration  #580\n",
      "-- Accuracy score : 0.8667560975609756\n",
      "Iteration  #590\n",
      "-- Accuracy score : 0.867390243902439\n",
      "Iteration  #600\n",
      "-- Accuracy score : 0.8679756097560976\n",
      "Iteration  #610\n",
      "-- Accuracy score : 0.8686341463414634\n",
      "Iteration  #620\n",
      "-- Accuracy score : 0.8694390243902439\n",
      "Iteration  #630\n",
      "-- Accuracy score : 0.8704146341463415\n",
      "Iteration  #640\n",
      "-- Accuracy score : 0.8712926829268293\n",
      "Iteration  #650\n",
      "-- Accuracy score : 0.8718292682926829\n",
      "Iteration  #660\n",
      "-- Accuracy score : 0.8724146341463415\n",
      "Iteration  #670\n",
      "-- Accuracy score : 0.8731219512195122\n",
      "Iteration  #680\n",
      "-- Accuracy score : 0.8737317073170732\n",
      "Iteration  #690\n",
      "-- Accuracy score : 0.8745121951219512\n",
      "Iteration  #700\n",
      "-- Accuracy score : 0.8751463414634146\n",
      "Iteration  #710\n",
      "-- Accuracy score : 0.8755853658536585\n",
      "Iteration  #720\n",
      "-- Accuracy score : 0.8758780487804878\n",
      "Iteration  #730\n",
      "-- Accuracy score : 0.876390243902439\n",
      "Iteration  #740\n",
      "-- Accuracy score : 0.8770243902439024\n"
     ]
    }
   ],
   "source": [
    "W_1, b_1, W_2, b_2, loss_history, ac_history = gradiant_decent(x_train, y_train, 750, 0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2b599f1b",
   "metadata": {},
   "source": [
    "## Part 4 - Plot the Loss and Accuracy over time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "7c9374f8",
   "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": [
    "# PLOT 1 - Plot the Loss \n",
    "\n",
    "# a - setup the plot\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "# b - plot loss \n",
    "ax.plot(loss_history, 'r')\n",
    "\n",
    "# c - labels and title\n",
    "ax.set_xlabel('Epoch')\n",
    "ax.set_ylabel('Loss')\n",
    "ax.set_title('Loss over time')\n",
    "\n",
    "# d - display \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "cbd957cc",
   "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": [
    "# PLOT 2 - Plot the Accuracy over time\n",
    "\n",
    "# a - setup the plot\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "# b - plot accuracy\n",
    "ax.plot(ac_history,  'b')\n",
    "\n",
    "# c - labels and title\n",
    "ax.set_xlabel('Epoch')\n",
    "ax.set_ylabel('Accuracy')\n",
    "ax.set_title('Accuracy over time')\n",
    "\n",
    "# d - display \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a7aaeb97",
   "metadata": {},
   "source": [
    "## Part 5 - Test performance on some examples, overall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "67c842ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_predictions(X, W_1, b_1, W_2, b_2):\n",
    "    _, _, _, A_2 = forward_propigation(W_1, b_1, W_2, b_2, X)\n",
    "    predictions = get_predictions(A_2)\n",
    "    return predictions\n",
    "\n",
    "def test_prediction(index, W_1, b_1, W_2, b_2):\n",
    "    current_image = x_train[:, index, None]\n",
    "    prediction = make_predictions(x_train[:, index, None], W_1, b_1, W_2, b_2)\n",
    "    label = y_train[index]\n",
    "    print(\"Prediction: \", prediction)\n",
    "    print(\"Label: \", label)\n",
    "    \n",
    "    current_image = current_image.reshape((28, 28)) * 255\n",
    "    plt.gray()\n",
    "    plt.imshow(current_image, interpolation='nearest')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "fac3eb9e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction:  [2]\n",
      "Label:  2\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction:  [2]\n",
      "Label:  2\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction:  [2]\n",
      "Label:  3\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction:  [4]\n",
      "Label:  4\n"
     ]
    },
    {
     "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": [
    "test_prediction(0, W_1, b_1, W_2, b_2)\n",
    "test_prediction(1, W_1, b_1, W_2, b_2)\n",
    "test_prediction(2, W_1, b_1, W_2, b_2)\n",
    "test_prediction(3, W_1, b_1, W_2, b_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "ea1e1a8b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.858"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Accuracy on Test Dataset\n",
    "test_predictions = make_predictions(x_test, W_1, b_1, W_2, b_2)\n",
    "get_accuracy(test_predictions, y_test)"
   ]
  }
 ],
 "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": "3.9.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}