ImageNeuralNetwork1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "4ff901cd-adee-4972-abfe-b8bbb2b29488",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a77cc658-4f24-4606-8dd2-57acebf99f13",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train=np.loadtxt('train_X.csv',delimiter=',').T\n",
    "Y_train=np.loadtxt('train_label.csv',delimiter=',').T\n",
    "\n",
    "X_test=np.loadtxt('test_X.csv',delimiter=',').T\n",
    "Y_test=np.loadtxt('test_label.csv',delimiter=',').T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8998e709-6c98-4ed1-9f32-602e1eb2c5c7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape of X_train:  (784, 1000)\n",
      "shape of Y_train:  (10, 1000)\n",
      "shape of X_test:  (784, 350)\n",
      "shape of X_train:  (10, 350)\n"
     ]
    }
   ],
   "source": [
    "print(\"shape of X_train: \",X_train.shape)\n",
    "print(\"shape of Y_train: \",Y_train.shape)\n",
    "print(\"shape of X_test: \",X_test.shape)\n",
    "print(\"shape of X_train: \",Y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "aa86ead3-241d-43ad-ad53-e25c89a530f2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x15d8d6190d0>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "index=random.randrange(0,X_train.shape[1])\n",
    "plt.imshow(X_train[:,index].reshape(28,28),cmap='gray')\n",
    "                                   "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "097dd5e8-1bac-4e16-9e6f-026f671f304d",
   "metadata": {},
   "source": [
    "## project"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "bef8d8cc-1871-4d28-baf0-8f4bfe8611bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "def tanh(x):\n",
    "    return np.tanh(x)\n",
    "\n",
    "def relu(x):\n",
    "    return np.maximum(x,0)\n",
    "\n",
    "def softmax(x):\n",
    "    expX=np.exp(x)\n",
    "    return expX/np.sum(expX,axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "97bb415f-f843-43f2-b445-c26d633d3aaf",
   "metadata": {},
   "outputs": [],
   "source": [
    "def derivative_tanh(x):\n",
    "    return (1-np.power(np.tanh(x),2))\n",
    "\n",
    "def derivative_relu(x):\n",
    "    return np.array(x>0,dtype=np.float32)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f41208d-6068-4242-abce-8d8b94f4a7dd",
   "metadata": {},
   "source": [
    "## intialization ther parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "190d47d9-8746-4b25-b2e4-bc62bfd1a98f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def initialize_parameters(n_x, n_h, n_y):\n",
    "    w1 = np.random.randn(n_h, n_x)*0.01\n",
    "    b1 = np.zeros((n_h, 1))\n",
    "    \n",
    "    w2 = np.random.randn(n_y, n_h)*0.01\n",
    "    b2 = np.zeros((n_y, 1))\n",
    "    \n",
    "    parameters = {\n",
    "        \"w1\" : w1,\n",
    "        \"b1\" : b1,\n",
    "        \"w2\" : w2,\n",
    "        \"b2\" : b2\n",
    "    }\n",
    "    \n",
    "    return parameters\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9020853c-0d99-4082-bca4-b5d7c704388e",
   "metadata": {},
   "source": [
    "## forward progation: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f6cbe991-7025-4216-9c8a-1e7e4a258350",
   "metadata": {},
   "outputs": [],
   "source": [
    "def forward_pro(x,parameters):\n",
    "    w1=parameters['w1']\n",
    "    b1=parameters['b1']\n",
    "    w2=parameters['w2']\n",
    "    b2=parameters['b2']\n",
    "\n",
    "    z1=np.dot(w1,x)+b1\n",
    "    a1=tanh(z1)\n",
    "\n",
    "    z2=np.dot(w2,a1)+b2\n",
    "    a2=softmax(z2)\n",
    "\n",
    "    forward_cache={\n",
    "        \"z1\" : z1,\n",
    "        \"a1\" : a1,\n",
    "        \"z2\" : z2,\n",
    "        \"a2\" : a2\n",
    "    }\n",
    "    return forward_cache\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c8c99ea-e061-4ad8-8424-5a3e96125015",
   "metadata": {},
   "source": [
    "## cost Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "79310148-6f2e-4fc3-bf78-61453eaa257a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def cost_function(a2,y):\n",
    "    m=y.shape[1]\n",
    "    cost=-(1/m)*np.sum(y*np.log(a2))\n",
    "    #cost= -(1/m)*np.sum(np.sum(y*np.log(a2,0),1))\n",
    "    return cost"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb03012a-6266-4972-8bf1-abc13758c902",
   "metadata": {},
   "source": [
    "## backward Propagation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b5a6cc67-0b1f-4129-99b0-43966bf03469",
   "metadata": {},
   "outputs": [],
   "source": [
    "def backward_pro(x,y,parameters,forward_cache):\n",
    "    w1=parameters['w1']\n",
    "    b1=parameters['b1']\n",
    "    w2=parameters['w2']\n",
    "    b2=parameters['b2']\n",
    "\n",
    "    a1=forward_cache['a1']\n",
    "    a2=forward_cache['a2']\n",
    "\n",
    "    m=x.shape[1]\n",
    "\n",
    "    dz2=(a2-y)\n",
    "    dw2=(1/m)*np.dot(dz2,a1.T)\n",
    "    db2=(1/m)*np.sum(dz2,axis=1,keepdims=True)\n",
    "\n",
    "    dz1=(1/m)*np.dot(w2.T,dz2)*derivative_relu(a1)\n",
    "    dw1=(1/m)*np.dot(dz1,x.T)\n",
    "    db1=(1/m)*np.sum(dz1,axis=1,keepdims=True)\n",
    "\n",
    "    gradients={\n",
    "        \"dw1\" : dw1,\n",
    "        \"db1\" : db1,\n",
    "        \"db2\" : db2,\n",
    "        \"dw2\" : dw2\n",
    "    }\n",
    "\n",
    "    return gradients\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c9835a2-6912-4127-a46a-50054ed54d39",
   "metadata": {},
   "source": [
    "## Upadate Parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "63c39580-ee62-4930-811b-d04d64b68376",
   "metadata": {},
   "outputs": [],
   "source": [
    "def update_parameters(parameters,gradients,learning_rate):\n",
    "\n",
    "    w1=parameters['w1']\n",
    "    b1=parameters['b1']\n",
    "    w2=parameters['w2']\n",
    "    b2=parameters['b2']\n",
    "\n",
    "    dw1=gradients['dw1']\n",
    "    db1=gradients['db1']\n",
    "    dw2=gradients['dw2']\n",
    "    db2=gradients['db2']\n",
    "\n",
    "    w1=w1-learning_rate*dw1\n",
    "    b1=b1-learning_rate*db1\n",
    "    w2=w2-learning_rate*dw2\n",
    "    b2=b2-learning_rate*db2\n",
    "\n",
    "    parameters={\n",
    "        \"w1\" : w1,\n",
    "        \"b1\" : b1,\n",
    "        \"w2\" : w2,\n",
    "        \"b2\" : b2\n",
    "    }\n",
    "\n",
    "    return parameters"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e2cce779-00cc-46ea-8cee-f9806bf210ac",
   "metadata": {},
   "source": [
    "## Complete Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f69b6039-4326-4369-b3b9-2c80d43bb0a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "def model(x,y,n_h,learning_rate,iterations):\n",
    "    n_x=x.shape[0]\n",
    "    n_y=y.shape[0]\n",
    "\n",
    "    cost_list=[]\n",
    "    \n",
    "    parameters=initialize_parameters(n_x,n_h,n_y)\n",
    "\n",
    "    for i in range(iterations):\n",
    "        forward_cache=forward_pro(x,parameters)\n",
    "        cost =cost_function(forward_cache['a2'],y)\n",
    "        gradients=backward_pro(x,y,parameters,forward_cache)\n",
    "        parameters=update_parameters(parameters,gradients,learning_rate)\n",
    "        cost_list.append(cost)\n",
    "        if(i%(iterations/10)==0):\n",
    "            print(\"cost after\",i,\"iterations is: \",cost)\n",
    "\n",
    "    return parameters,cost_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "310892fb-12f3-48c3-8484-1670fb7e2d39",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cost after 0 iterations is:  2.338116984902653\n",
      "cost after 10 iterations is:  2.1345268132648387\n",
      "cost after 20 iterations is:  1.9636033989235457\n",
      "cost after 30 iterations is:  1.8181170003462586\n",
      "cost after 40 iterations is:  1.693239244053185\n",
      "cost after 50 iterations is:  1.5851868186496558\n",
      "cost after 60 iterations is:  1.4909856203232592\n",
      "cost after 70 iterations is:  1.4083004803100811\n",
      "cost after 80 iterations is:  1.3352749074707286\n",
      "cost after 90 iterations is:  1.2704132142923503\n"
     ]
    }
   ],
   "source": [
    "iterations=100\n",
    "n_h=1000\n",
    "learning_rate=0.002\n",
    "Parameters,Cost_list=model(X_train,Y_train,n_h=n_h,learning_rate=learning_rate,iterations=iterations)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d6a41326-110a-4ec3-868f-0a2a1109184f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAGdCAYAAADAAnMpAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA+PUlEQVR4nO3dZ3hUZcLG8XsmPZAMJBCSkAChSQ+BgFQpCiqIskoRS0TlVVyasjbUVXd1xXXtoqisDakCAVERAUFCL4HQO4EkkNBJg9Q57wd0dllFk5DkZGb+v+s6H3LmHObOcyFzO+c5z7EYhmEIAADAJFazAwAAAPdGGQEAAKaijAAAAFNRRgAAgKkoIwAAwFSUEQAAYCrKCAAAMBVlBAAAmMrT7AAlYbfbdfz4cQUEBMhisZgdBwAAlIBhGMrOzlZ4eLis1it//+EUZeT48eOKjIw0OwYAACiD1NRURUREXPF1pygjAQEBki79MoGBgSanAQAAJZGVlaXIyEjH5/iVOEUZ+eXSTGBgIGUEAAAn80dTLJjACgAATEUZAQAApqKMAAAAU1FGAACAqSgjAADAVJQRAABgKsoIAAAwFWUEAACYijICAABMRRkBAACmoowAAABTUUYAAICp3LqM7DyWqbumrNfZ3AKzowAA4LbctowYhqEn527X2kNn9Oz8HTIMw+xIAAC4JbctIxaLRa8NaiNPq0Xf78zQ10nHzY4EAIBbctsyIkmt6to09vomkqTnv96p9MyLJicCAMD9uHUZkaQ/92yk6Mgaysor0pNzt3O5BgCASub2ZcTTw6o3BkfLx9OqVQdOa9qGFLMjAQDgVty+jEhS45DqeuqmZpKkV77boyOnc01OBACA+6CM/Gx4lwbq3DBYFwuLNf6rJBUV282OBACAW6CM/Mxqtehfg9sowMdTW1LO68OVh8yOBACAW6CM/JeImv568daWkqS3lx3Q9rTz5gYCAMANUEb+x+3t6qpf61AV2Q09OjtJFwuKzY4EAIBLo4z8D4vFon8MbK2QAB8dPpWrid/vMTsSAAAujTLyG2pW89brg6MlSVPXHdVP+06anAgAANdFGbmC65rW1vAuDSRJT8zdzsP0AACoIJSR3/H0zc3UOKS6TmXna0I8q7MCAFARKCO/w9fLQ28PbSsvD4t+2HVCszalmh0JAACXQxn5A63q2vTEjddIkv7+zW4dOpVjciIAAFwLZaQERnRrqK6NL63O+uisJBUUsTorAADlhTJSAlarRW8Mbqsa/l7acSxTbyzdZ3YkAABcBmWkhEJtvnr19jaSpI8TDmvtwdMmJwIAwDVQRkrhplahGtYxUoYhjf9qm85xuy8AAFeNMlJKf72lhRrWqqaMrDw9OY/bfQEAuFqUkVLy9/bUu8Ni5O1h1dLdJzRt/VGzIwEA4NQoI2XQqq5NT93cTJL00nd7tDcjy+REAAA4L8pIGT3QtYF6NwtRQZFdY2Zs5em+AACUEWWkjCwWi/41qI1qB/jowMkcvfTdbrMjAQDglCgjVyG4uo/eHtpWFos0Y0OKvt+RbnYkAACcDmXkKnVtXEuP9GgkSXpy3nalnr1gciIAAJwLZaQcPNanqdrVq6HsvCKNnrmV5eIBACgFykg58PKw6t1hMQr09dS21PN6fQnLxQMAUFKUkXISUdNf/xocLenScvHL954wOREAAM6BMlKObmwZquFdGkiS/vLVNqVnXjQ3EAAAToAyUs4m9GumVnUDde5CocbNTFJRMfNHAAD4PZSRcubj6aFJw9qpuo+nNh45q7eW7Tc7EgAAVRplpAI0qFVNE29vLUl6f8Uh/bTvpMmJAACouigjFWRAdLju6VRPkvTY7CQdP8/8EQAAfgtlpAI917+FWoZfmj8yZuZWFTJ/BACAX6GMVCBfLw99cHc7Bfh4KvHoOb3+A+uPAADwvygjFax+cDW9NqiNJOmjhMNatpv1RwAA+G+lKiMTJ05Uhw4dFBAQoJCQEA0cOFD79v3+/+3Hx8erT58+ql27tgIDA9W5c2f98MMPVxXa2dzcOsyx/sj4r5J4fg0AAP+lVGVk5cqVGjVqlNavX6+lS5eqqKhIffv2VW5u7hXPSUhIUJ8+fbRo0SIlJiaqV69eGjBggLZu3XrV4Z3JM/2aq21kDWXlFemR6YnKKyw2OxIAAFWCxTAMo6wnnzp1SiEhIVq5cqWuu+66Ep/XsmVLDR06VM8//3yJjs/KypLNZlNmZqYCAwPLGtd0x85f1C3vrtK5C4W669p6euVPrc2OBABAhSnp5/dVzRnJzMyUJAUFBZX4HLvdruzs7N89Jz8/X1lZWZdtrqBuDT+9fWeMLBZpxoYUxW9JMzsSAACmK3MZMQxD48ePV7du3dSqVasSn/fGG28oNzdXQ4YMueIxEydOlM1mc2yRkZFljVnl9GhaW2N7N5EkPTN/h/ZmuEbRAgCgrMpcRkaPHq3t27dr5syZJT5n5syZevHFFzV79myFhIRc8bgJEyYoMzPTsaWmppY1ZpU09vom6t6klvIK7frztC3Kzis0OxIAAKYpUxkZM2aMFi5cqBUrVigiIqJE58yePVsPPvigvvrqK91www2/e6yPj48CAwMv21yJh9Wid+6MUbjNV4dP5+qJOdt1FVN3AABwaqUqI4ZhaPTo0YqPj9fy5csVFRVVovNmzpyp4cOHa8aMGerfv3+ZgrqaoGre+uCe9vL2sGrxrgx9lHDY7EgAAJiiVGVk1KhRmjZtmmbMmKGAgABlZGQoIyNDFy/+57krEyZMUFxcnOPnmTNnKi4uTm+88YY6derkOOeXya/urG1kDb1wawtJ0muL92rNwdMmJwIAoPKVqoxMnjxZmZmZ6tmzp8LCwhzb7NmzHcekp6crJSXF8fNHH32koqIijRo16rJzxo0bV36/hRO7q2M9DW4fIbshjZm5lQfqAQDczlWtM1JZXGWdkSvJKyzWHZPXatfxLEVH1tBXD3eSj6eH2bEAALgqlbLOCMqHr5eHPrynvWx+XtqWel4vLtxtdiQAACoNZaSKiAzy1zt3tpXFIs3cmKKZG1P++CQAAFwAZaQK6XlNiP7Sp6kk6YWvd2lLyjmTEwEAUPEoI1XMn3s21o0t66ig2K5HpiXqZHae2ZEAAKhQlJEqxmq16I0hbdU4pLpOZOVr1PQtKiiymx0LAIAKQxmpgqr7eOrje9srwMdTm46c0z++Y0IrAMB1UUaqqIa1q+vtO9tKkr5Yd1RfbXat5/MAAPALykgVdn3zOnrshksTWp+bv1NbmdAKAHBBlJEqbkzv/0xoffjLRJ3IYkIrAMC1UEaquF8mtDatU10ns/P18JeJyissNjsWAADlhjLiBKr7eGpKXKxsfl5KSj2v5xbslBOs4g8AQIlQRpxE/eBqev+udrJapLmJafp87RGzIwEAUC4oI06kW5NaeqZfc0nSy9/t0eoDp01OBADA1aOMOJkHu0Xp9nZ1VWw39OfpiUo+nWt2JAAArgplxMlYLBa98qfWiqlXQ1l5RRrxxSZl5RWaHQsAgDKjjDghXy8PfXRve4XZfHXoVK7GzNiqYjsTWgEAzoky4qRCAnw1JS5Wvl5Wrdx/Sq9+v8fsSAAAlAllxIm1qmvT64OjJUlTViVrDkvGAwCcEGXEyd3SJlxjezeWJD0zf4c2Jp81OREAAKVDGXEBj97QVP1ah6qw2NDDX25WypkLZkcCAKDEKCMuwGq16I3BbdUmwqZzFwr1AHfYAACcCGXERfh5e2hKXKxCA3118GSORs/YqqJiu9mxAAD4Q5QRF1In0Ff/vi9Wfl4eSth/Si9/xx02AICqjzLiYlrVtemtoZfusPl87RF9wTNsAABVHGXEBd3UKkxP3nSNJOlv3+zSir0nTU4EAMCVUUZc1CM9GmlIbITshjR6xhbtSc8yOxIAAL+JMuKiLBaLXh7YWp0bBiu3oFgPfr5JJ7PyzI4FAMCvUEZcmLenVR/e014Na1fT8cw8jZi6WRcLis2OBQDAZSgjLs7m76XPhndQTX8vbU/L1LhZPFQPAFC1UEbcQP3gavo4LlbeHlYt2X1Cryzill8AQNVBGXETHRoE6fUhl275/WR1sj5fk2xyIgAALqGMuJFbo8Mdt/z+/dvdWrr7hMmJAACgjLidR3o00rCO9WQ3pDEzt2hb6nmzIwEA3BxlxM1YLBa9dFtL9WhaW3mFdj34xWalnuUpvwAA81BG3JCnh1Xv391OzcMCdTonX/d9tlHnLxSYHQsA4KYoI26quo+nPr+/g8Jtvjp8KlcjvtisvELWIAEAVD7KiBurE+irzx/oqABfT20+ek6PzU6SnTVIAACVjDLi5prWCdDH915ag+T7nRl6+TvWIAEAVC7KCNS5UbBjDZJP1yTr36sOm5wIAOBOKCOQdGkNkmf6NZMkvfzdHn2ddMzkRAAAd0EZgcP/dW+o+7s2kCQ9Pmeb1hw8bW4gAIBboIzAwWKx6K/9W6h/mzAVFht6+MtE7TqeaXYsAICLo4zgMlarRW8OiVanhkHKyS/S8M82sSgaAKBCUUbwKz6eHvo4LlbNQgN0KjtfcZ9u1JmcfLNjAQBcFGUEvynQ10tfPNBRdWv4Kfl0ru7/fJNy8ovMjgUAcEGUEVxRnUBfTX2wo2r6e2l7WqZGfpmogiK72bEAAC6GMoLf1ah2dX12f0f5e3to9cHTGv8Vq7QCAMoXZQR/qG1kDX14T3t5eVj07fZ0/e2bXTIMCgkAoHxQRlAi1zWtrdcHR8tikb5Yd1TvLT9odiQAgIsoVRmZOHGiOnTooICAAIWEhGjgwIHat2/fH563cuVKtW/fXr6+vmrYsKE+/PDDMgeGeW5rW1cv3NJCkvTm0v36cv1RkxMBAFxBqcrIypUrNWrUKK1fv15Lly5VUVGR+vbtq9zc3Cuek5ycrH79+ql79+7aunWrnnnmGY0dO1bz5s276vCofMO7Rmls78aSpOe/3smy8QCAq2YxruLi/6lTpxQSEqKVK1fquuuu+81jnnrqKS1cuFB79vznabAjR47Utm3btG7duhK9T1ZWlmw2mzIzMxUYGFjWuCgnhmHohYW7NHXdUXlaLZoSF6tezULMjgUAqGJK+vl9VXNGMjMvLRUeFBR0xWPWrVunvn37Xrbvxhtv1ObNm1VYWPib5+Tn5ysrK+uyDVWHxWLRiwNa6ra24SqyG3pkeqI2HTlrdiwAgJMqcxkxDEPjx49Xt27d1KpVqysel5GRoTp16ly2r06dOioqKtLp07/9ILaJEyfKZrM5tsjIyLLGRAWxWi16fXC0el1TW3mFdj3w+SaeYwMAKJMyl5HRo0dr+/btmjlz5h8ea7FYLvv5lytD/7v/FxMmTFBmZqZjS01NLWtMVCAvD6s+uLu9OjSoqey8It336UYdOpVjdiwAgJMpUxkZM2aMFi5cqBUrVigiIuJ3jw0NDVVGRsZl+06ePClPT08FBwf/5jk+Pj4KDAy8bEPV5OftoU+Gd1DL8ECdzinQPf/eoLRzPFgPAFBypSojhmFo9OjRio+P1/LlyxUVFfWH53Tu3FlLly69bN+SJUsUGxsrLy+v0qVFlRTo66WpD3RUo9rVlJ6Zp3v+vUEns/PMjgUAcBKlKiOjRo3StGnTNGPGDAUEBCgjI0MZGRm6ePGi45gJEyYoLi7O8fPIkSN19OhRjR8/Xnv27NGnn36qTz75RI8//nj5/RYwXXB1H00bca0iavrpyJkLivtko85fKDA7FgDACZSqjEyePFmZmZnq2bOnwsLCHNvs2bMdx6SnpyslJcXxc1RUlBYtWqSffvpJbdu21UsvvaR3331Xd9xxR/n9FqgSwmx+mj7iWoUE+GhvRrbu+4wn/QIA/thVrTNSWVhnxLnsP5GtIR+t0/kLheoYFaQv7u8oP28Ps2MBACpZpawzAvyWpnUC9OUD1yrAx1Mbk8/qoS83K7+o2OxYAIAqijKCCtE6wqbP7u8gPy8PrTpwWqNnbFVhsd3sWACAKogyggoT2yBI/74vVt6eVi3dfUKPzU5Ssb3KXxUEAFQyyggqVNfGtfThPe3k5WHRt9vT9dS87bJTSAAA/4UyggrXu1kdvXtnjDysFs1NTNNzX++UE8ybBgBUEsoIKsXNrcP05pBoWS3SjA0p+ts3uykkAABJlBFUotva1tVrg6JlsUifrz2iVxbtoZAAACgjqFyD2kfolT+1liRNWZWs15fso5AAgJujjKDSDetYT3+/raUk6f0Vh/TWsgMmJwIAmIkyAlPEdW6g5/o3lyS9++MBvUMhAQC3RRmBaUZ0b6hn+jWTJL21bL8mLaeQAIA7oozAVA9d10hP3XSpkLy+ZL/eX3HQ5EQAgMpGGYHpHunZSE/ceI0k6V8/7NPknw6ZnAgAUJkoI6gSRvVqrPF9mkqS/rl4rz74iW9IAMBdUEZQZYy9vomjkLy2eB+XbADATVBGUKWMvb6J/vJzIfnXD/uY1AoAboAygipnzPVN9HjfS4Xk9SX79d6PFBIAcGWUEVRJo3s3cUxqfWPpfr29bD8rtQKAi6KMoMoa1aux47bft5cd0BtLKCQA4IooI6jSHunZyLFS66QVB/Xq4r0UEgBwMZQRVHkjujfUCwNaSJI+WnlYL33L034BwJVQRuAU7u8apZcGtpIkfbomWS8s3CW7nUICAK6AMgKncW+n+pp4e2tZLNLUdUc1IX6HiikkAOD0KCNwKsM61tPrg6JltUizN6fqL18lqajYbnYsAMBVoIzA6dzRPkLvDouRp9WiBUnHNWbmVhUUUUgAwFlRRuCUbmkTrg/ubidvD6u+35mhR6YlKq+w2OxYAIAyoIzAafVtGaop98XKx9OqH/ee1INfbFJufpHZsQAApUQZgVPr0bS2Pr+/o6p5e2jNwTOK+3SjMi8Wmh0LAFAKlBE4vc6NgjVtxLUK9PVU4tFzumvKep3JyTc7FgCghCgjcAkx9Wpq1kOdFVzNW7uOZ2nox+t1IivP7FgAgBKgjMBltAgP1FcjOyvM5quDJ3M06MO1SjlzwexYAIA/QBmBS2lUu7q+eriz6gX5K/XsRQ36cK32ZWSbHQsA8DsoI3A5kUH+mjuys66pE6CT2fka8tE6bU05Z3YsAMAVUEbgkkICfTX74U5qG1lDmRcLdfe/N2jNwdNmxwIA/AbKCFxWDX9vTR9xrbo2DtaFgmLd/9kmLd6ZYXYsAMD/oIzApVXz8dSnwzvoxpZ1VFBs15+nJ+qrTalmxwIA/BfKCFyej6eH3r+rnYbERshuSE/O264PVx4yOxYA4GeUEbgFTw+r/nlHGz3co6Ek6dXv92rioj0yDMPkZAAAygjchsVi0YSbm2vCzc0kSR8lHNZT87arqJgn/gKAmSgjcDsP92ik1+5oI6tF+mpzmkZOS9TFAp74CwBmoYzALQ3pEKkP72kvH0+rlu05qXs/2aDzFwrMjgUAbokyArfVt2Wovnzw0gP2Nh89pyEfrVN65kWzYwGA26GMwK11jArSVyM7q06gj/afyNEdH6zVwZMsHw8AlYkyArfXLDRQ8x7pooa1q+l4Zp7umLxOm4+cNTsWALgNygggKaKmv+aO7KKYev9ZPv6HXazWCgCVgTIC/CyomrdmjOikG5qHKL/IrkemJerL9UfNjgUALo8yAvwXP28PfXhPew3rGCm7If11wU7964e9LI4GABWIMgL8D08Pq175U2s9dkNTSdL7Kw7pL19tU0ERi6MBQEWgjAC/wWKxaNwNTfTPO1rLw2pR/NZjuv/zjcrKKzQ7GgC4nFKXkYSEBA0YMEDh4eGyWCxasGDBH54zffp0RUdHy9/fX2FhYbr//vt15syZsuQFKtXQDvX0yX2xqubtoTUHz2jIh6xFAgDlrdRlJDc3V9HR0Zo0aVKJjl+9erXi4uL04IMPateuXZozZ442bdqkESNGlDosYIae14Ro9sOdVTvAR3szsvWn99dqT3qW2bEAwGWUuozcfPPNevnll3X77beX6Pj169erQYMGGjt2rKKiotStWzc9/PDD2rx5c6nDAmZpVdem+X/uosYh1ZWRlafBH65Twv5TZscCAJdQ4XNGunTporS0NC1atEiGYejEiROaO3eu+vfvf8Vz8vPzlZWVddkGmC2ipr/mjeyiTg2DlJNfpPs/36SZG1PMjgUATq9Sysj06dM1dOhQeXt7KzQ0VDVq1NB77713xXMmTpwom83m2CIjIys6JlAiNn8vTX3gWt0eU1fFdkMT4nfon4v3ym7n1l8AKKsKLyO7d+/W2LFj9fzzzysxMVGLFy9WcnKyRo4cecVzJkyYoMzMTMeWmppa0TGBEvP2tOqNIdEad30TSdLknw5p7KytyissNjkZADgni3EVqzlZLBbNnz9fAwcOvOIx9957r/Ly8jRnzhzHvtWrV6t79+46fvy4wsLC/vB9srKyZLPZlJmZqcDAwLLGBcrd3MQ0PT1vu4rshtrVq6EpcbEKru5jdiwAqBJK+vld4d+MXLhwQVbr5W/j4eEhSaxqCac3qH2Epj7YUYG+ntqScl4DP1ijAyd46i8AlEapy0hOTo6SkpKUlJQkSUpOTlZSUpJSUi5N5JswYYLi4uIcxw8YMEDx8fGaPHmyDh8+rDVr1mjs2LHq2LGjwsPDy+e3AEzUpVEtxf+5q+oH+yv17EXdPnmtVh84bXYsAHAapS4jmzdvVkxMjGJiYiRJ48ePV0xMjJ5//nlJUnp6uqOYSNLw4cP15ptvatKkSWrVqpUGDx6sa665RvHx8eX0KwDmaxxSXfP/3FWx9WsqO69I9322kTttAKCErmrOSGVhzgicRX5RsZ6au10Lko5Lkh7sFqVn+jWXh9VicjIAqHxVZs4I4E58PD301tC2Gt/n0kP2PlmdrP+bulnZPNMGAK6IMgKUM4vForHXN9Gku2Lk42nV8r0nNWjyOqWevWB2NACokigjQAW5pU2445k2+05ka+D7a7T5yFmzYwFAlUMZASpQ28gaWji6q1qEBepMboGGTVmvOZtZxA8A/htlBKhgYTY/zX2ks25qGarCYkNPzN2uf3y3W8UsIQ8AkigjQKXw9/bUB3e309ifl5CfsipZI77YpCwmtgIAZQSoLFarReP7NNV7wy5NbF2x75Ru/2Ctkk/nmh0NAExFGQEq2YDocM0d2UWhgb46eDJHt01arYT9p8yOBQCmoYwAJmgdYdPC0V3Vrl4NZeUVafhnG/XvVYd5XhMAt0QZAUwSEuirmQ910pDYCNkN6eXv9ugvc7Ypr7DY7GgAUKkoI4CJfDw99M872uiFAS3kYbUofssxDf1ondIzL5odDQAqDWUEMJnFYtH9XaM09YGOquHvpW1pmRrw3mptYoE0AG6CMgJUEV0b19I3o7upWWiATucUaNjH6/Xl+qPMIwHg8igjQBUSGeSv+D930S1twlRkN/TXBTv19LwdzCMB4NIoI0AV4+/tqfeGxejpm5vJapFmb07V0I/X6/h55pEAcE2UEaAKslgsGtmjkT6//+d5JKnnNeC91Vp36IzZ0QCg3FFGgCrsuqa19c3obo4H7d3zyQbWIwHgcigjQBUXGeSveY900Z9i6qrYbujl7/Zo7Kwk5eYXmR0NAMoFZQRwAn7eHnpzSLReGNBCnlaLvtl2XH/6YI0On8oxOxoAXDXKCOAkflmPZOZDnRQS4KP9J3J066Q1Wrwzw+xoAHBVKCOAk+nQIEjfju2mjlFByskv0shpiXr1+70qKrabHQ0AyoQyAjihkABfTR9xrUZ0i5IkfbjykO75ZINOZueZnAwASo8yAjgpLw+rnrulhd6/q52qeXto/eGzuuXd1dqYzDLyAJwLZQRwcv3bhGnhmG5qWqe6Tmbna9iU9fo44RC3/wJwGpQRwAU0ql1dC0Z11cC24Sq2G3pl0V499GWiMi8Wmh0NAP4QZQRwEf7ennpraFu9NLCVvD2sWrr7hG55b5V2pGWaHQ0AfhdlBHAhFotF93aqr3mPdFFkkJ9Sz17UHZPX6st1R7hsA6DKoowALqh1hE3fjumuvi3qqKDYrr9+vUtjZyUpO4/LNgCqHsoI4KJsfl766N72eq5/c8eqrQPeW62dx7hsA6BqoYwALsxisWhE94aa/XBnhdt8deTMBd0+ea2+XH+UyzYAqgzKCOAG2tevqUXjuuuG5iEqKLLrrwt2avSMrcrisg2AKoAyAriJGv7emhIX67hs892OdN3y7mptSz1vdjQAbo4yAriRXy7bzBnZWXVr+Cnl7AXdMXmtpiQclt3OZRsA5qCMAG4opt6lyzY3twpVkd3QPxbt0YNfbNKZnHyzowFwQ5QRwE3Z/Lz0wd3t9PLAVvL2tGrFvlPq9+4qrT102uxoANwMZQRwYxaLRfd0qq+vR3VVo9rVdCIrX3f/e4P+9cNeFRbbzY4HwE1QRgCoeVigvhnTTXd2iJRhSO+vOKQhH61T6tkLZkcD4AYoIwAkXXq2zat3tNH7d7VTgK+ntqacV793VmnhtuNmRwPg4igjAC7Tv02Yvh/XXe3r11R2fpHGztyqv3y1TTn5RWZHA+CiKCMAfiWipr9mP9RJY3s3ltUizduSpv7vrlISa5IAqACUEQC/ydPDqvF9r9Gshy6tSXL0zKU1SSYtP6Bi1iQBUI4oIwB+V8eoIC0a1123tAlTsd3Q60v2a9jH65ncCqDcUEYA/CGbn5feGxaj1wdHq5q3hzYeOat+76zSgq3HzI4GwAVQRgCUiMVi0aD2Efp+3HVqV6+GsvOL9OjsJI2duVWZF3ngHoCyo4wAKJV6wf766uHOGt+nqTysFi3cdlw3vZ2gtQdZuRVA2VBGAJSap4dVY69vorkjO6tBsL/SM/N017836KVvdyuvsNjseACcDGUEQJnF1Kup78Z217CO9SRJn6xO1q2TVmvX8UyTkwFwJpQRAFelmo+nJt7eWp/cF6ta1b21/0SOBr6/Rh/8dJBbgAGUCGUEQLm4vnkd/fDoderToo4Kiw29tnifhny0TkdO55odDUAVV+oykpCQoAEDBig8PFwWi0ULFiz4w3Py8/P17LPPqn79+vLx8VGjRo306aefliUvgCosuLqPPr63vf41qI2q+3gq8eg53fzOKn25/qgMg29JAPy2UpeR3NxcRUdHa9KkSSU+Z8iQIfrxxx/1ySefaN++fZo5c6aaNWtW2rcG4AQsFosGx0Zq8aPd1blhsC4WFuuvC3Yq7tONSs+8aHY8AFWQxbiK/12xWCyaP3++Bg4ceMVjFi9erDvvvFOHDx9WUFBQmd4nKytLNptNmZmZCgwMLGNaAJXNbjf0+doj+ufivcovsivA11MvDmip29vVlcViMTsegApW0s/vCp8zsnDhQsXGxuq1115T3bp11bRpUz3++OO6ePHK/4eUn5+vrKysyzYAzsdqteiBblH6bmx3tY2soey8Iv1lzjb939REnczOMzsegCqiwsvI4cOHtXr1au3cuVPz58/X22+/rblz52rUqFFXPGfixImy2WyOLTIysqJjAqhAjUOqa+7Iznrixmvk5WHRsj0ndONbCfpm23HmkgCo+Ms0ffv21apVq5SRkSGbzSZJio+P16BBg5Sbmys/P79fnZOfn6/8/HzHz1lZWYqMjOQyDeAC9qRn6S9fbdPu9EvfeN7cKlQvDWylWtV9TE4GoLxVmcs0YWFhqlu3rqOISFLz5s1lGIbS0tJ+8xwfHx8FBgZetgFwDc3DArVgVFeNu76JPK0Wfb8zQ33eXKlvtx83OxoAk1R4GenatauOHz+unJwcx779+/fLarUqIiKiot8eQBXk7WnVY32a6uvRXdUsNEDnLhRq9IytemRaok5l5//xHwDApZS6jOTk5CgpKUlJSUmSpOTkZCUlJSklJUWSNGHCBMXFxTmOv+uuuxQcHKz7779fu3fvVkJCgp544gk98MADv3mJBoD7aBlu08LR3TT2v74l6fvWSn2ddIy5JIAbKXUZ2bx5s2JiYhQTEyNJGj9+vGJiYvT8889LktLT0x3FRJKqV6+upUuX6vz584qNjdXdd9+tAQMG6N133y2nXwGAM/P2tGp8n6ZaMKqrmocF6tyFQo2blaT/m5qoE1nccQO4g6uawFpZWGcEcA+FxXZN/umQ3lt+QIXFhgJ9PfXcLS00uH0E65IATqjKTGAFgJLy8rBq7PVN9M2YbmoTYVNWXpGenLtdcZ9uVOrZC2bHA1BBKCMAqpxmoYGKf6SLnr65mXw8rVp14LRufDtBn69Jlp0nAQMuhzICoEry9LBqZI9G+n5cd3VsEKQLBcV68ZvdGvzROh08mW12PADliDICoEprWLu6Zj3USS8NbKVq3h5KPHpO/d5ZrXeWHVBBkd3seADKAWUEQJVntVp0b6f6Wjq+h3o3C1FBsV1vLduvW95bpcSj58yOB+AqUUYAOI3wGn765L5YvTcsRsHVvLX/RI4GfbhWL3y9U9l5hWbHA1BGlBEATsVisWhAdLiWje+hQe0jZBjSF+uOqs+bCVqyK8PseADKgDICwCnVrOat1wdHa9qD16pekL8ysvL00JeJGvkli6UBzoYyAsCpdWtSSz88ep0e6dlIHlaLFu/K0A1vrNSX646omNuAAadAGQHg9Py8PfTUTc307Zhuio6soez8Iv316126Y/Ja7T6eZXY8AH+AMgLAZTQPu7RY2t9ubanqPp5KSj2vAZNW65VFe3ShoMjseACugDICwKV4WC26r0sDLRvfQ/1ah6rYbujjhMPq82aClu0+YXY8AL+BMgLAJYXafPXB3e316fBY1a3hp2PnL2rE1M16aOpmHTt/0ex4AP4LZQSAS+vdrI6Wjr9OI3s0kqfVoiW7T+iGN1bqo5WHVFjMCq5AVUAZAeDy/L099fTNzbTo5+fcXCws1sTv96r/u6u04fAZs+MBbo8yAsBtNK0ToNkPd9Jrg9oo6OcVXId+vF7jZyfpVHa+2fEAt0UZAeBWLBaLhsRGavlfeuiua+vJYpHitx5T7zd+0lTWJgFMYTEMo8r/l5eVlSWbzabMzEwFBgaaHQeAC0lKPa/nFuzQzmOX1iNpGR6ov9/WSu3r1zQ5GeD8Svr5TRkB4PaK7Yambziq13/Yp6y8S+uRDG4foadubqZa1X1MTgc4r5J+fnOZBoDb87BaFNe5gZY/3lOD20dIkuYkpqn36z/pi7VHVMRdN0CF4psRAPgfiUfP6a8Ldmp3+qVLN81CA/T321qpY1SQyckA58JlGgC4CsV2QzM2HNXrS/Yr82KhJOm2tuF6pl9z1Qn0NTkd4BwoIwBQDs7mFuj1Jfs0c2OKDEPy9/bQ6N6N9WC3KPl4epgdD6jSKCMAUI52HsvU81/v1JaU85Kk+sH++mv/Frq+eYgsFou54YAqijICAOXMMAwtSDqmiYv26uTPi6Rd17S2nr+lhRqHVDc5HVD1UEYAoILk5Bfp/RUH9cmqZBUU2+X58904465vIpu/l9nxgCqDMgIAFezI6Vy9/N1uLdtzUpJU099Lf+l7jYZ1rCcPK5duAMoIAFSShP2n9NK3u3XgZI6kS7cCP39LC3VpXMvkZIC5KCMAUIkKi+2avv6o3lp2wHErcJ8WdfRMv+aKqlXN5HSAOSgjAGCCc7kFeufHA/py/VEV2w15eVh0X+cGGnN9E9n8mE8C90IZAQATHTyZrZe/26Of9p2SdGk+yaM3NNVd19aTlwdP4oB7oIwAQBXw076Tevm7PTr483yShrWr6dl+zdW7GeuTwPVRRgCgiigqtmvmplS9vXS/zuQWSJK6NArWs/2bq2W4zeR0QMWhjABAFZOVV6gPVhzSp6svrU9isUi3x0To8RubKszmZ3Y8oNxRRgCgiko9e0H/+mGfFm47Lkny9bJqRLeGGtmzkar7eJqcDig/lBEAqOKSUs/rle/2aOORs5Kk4GreevSGJrqzI5Nc4RooIwDgBAzD0JLdJ/Tq93uVfDpXktSwVjU9eVMz3diyDpNc4dQoIwDgRAqL7Zq1MUVvLzvgmOTavn5NTbi5mWIbBJmcDigbyggAOKHsvEJNSTisKauSdbGwWNKllVyfuukaNQ4JMDkdUDqUEQBwYiey8vT2sv2avSlVdkOyWqQhsZF69IamCrX5mh0PKBHKCAC4gIMnc/Ta4r1asvuEpEt33gzvEqVHejSSzZ/l5VG1UUYAwIUkHj2riYv2avPRc5KkQF9PPdKzsYZ3aSA/bw+T0wG/jTICAC7GMAwt33tSry3ep30nsiVJdQJ9NPb6JhoSG8ntwKhyKCMA4KKK7YYWbD2mN5fu17HzFyVJDYL99VifphrQJlxWK7cDo2qgjACAi8svKtb09Sl6f8VBx+3AzcMC9cSNTdXrGh7EB/NRRgDATeTkF+nT1cmaknBY2flFki6tUfJ432vUuVGwyengzigjAOBmzuUW6MOVh/T52iPKL7JLkro3qaXH+16j6Mga5oaDW6KMAICbOpGVp0nLD2rWphQVFl/6J75Pizoa36epmofxbygqD2UEANxc6tkLeufHA4rfkib7z//S928TpsduaMJqrqgUJf38LvV9YAkJCRowYIDCw8NlsVi0YMGCEp+7Zs0aeXp6qm3btqV9WwBAKUUG+ev1wdFa8lgPDYgOlyR9tz1dfd9K0PjZSTry84P5ALOVuozk5uYqOjpakyZNKtV5mZmZiouL0/XXX1/atwQAXIXGIdX13rAYfT+uu/q2qCO7IcVvPabr31ypJ+duU+rZC2ZHhJu7qss0FotF8+fP18CBA//w2DvvvFNNmjSRh4eHFixYoKSkpBK/D5dpAKD8bE87r7eW7teKfackSZ5WiwbHRmhUr8aKqOlvcjq4kgq7TFMWn332mQ4dOqQXXnihRMfn5+crKyvrsg0AUD7aRNTQZ/d3VPyfu6h7k1oqshuauTFVvV7/Sc/M3+FYSA2oLBVeRg4cOKCnn35a06dPl6enZ4nOmThxomw2m2OLjIys4JQA4H7a1aupLx+8VnNGdlaXRsEqLDY0Y0OKev5rhZ6llKASVWgZKS4u1l133aW//e1vatq0aYnPmzBhgjIzMx1bampqBaYEAPfWoUGQZvxfJ81+qJOjlEz/uZQ8M3+H0s4xpwQVq0LnjJw/f141a9aUh8d/nihpt9tlGIY8PDy0ZMkS9e7d+w/fhzkjAFB51h8+o7eX7df6w2clSV4eFg1qH6E/92ysyCDmlKDkSvr5XbLrJmUUGBioHTt2XLbvgw8+0PLlyzV37lxFRUVV5NsDAMqgU8NgzXqoszYcPqN3fjygtYfOaObGVM3ZnKbb29XVqF6NVT+4mtkx4UJKXUZycnJ08OBBx8/JyclKSkpSUFCQ6tWrpwkTJujYsWOaOnWqrFarWrVqddn5ISEh8vX1/dV+AEDVcm3DYM1oGKxNR87q3R8PaNWB0/pqc5rmbTmm26LDNap3YzWqXd3smHABpZ4zsnnzZsXExCgmJkaSNH78eMXExOj555+XJKWnpyslJaV8UwIATNOhQZC+fPBazXuki3peU1vFdkPxW4/phjdXavSMLdqbwR2PuDosBw8AKJXtaef17o8HtWzPCce+vi3qaHTvxmoTUcO8YKhyeDYNAKBC7TqeqQ9WHNKinen65ZOkR9PaGt27sTo0CDI3HKoEyggAoFIcPJmtD346pK+Tjqv45yfydYwK0qhejXVdk1qyWCwmJ4RZKCMAgEqVcuaCJq88pHmJaSootkuSWtUN1KiejXVjy1BZrZQSd0MZAQCYIiMzT1NWHdaMDSm6WFgsSWpUu5pG9mik29rWlbdnpTyJBFUAZQQAYKqzuQX6bE2yPl97RNl5RZKkcJuvRnRvqDs7Rsrfu0KXukIVQBkBAFQJ2XmFmr4hRZ+sTtap7HxJUk1/Lw3vEqW4zvVVs5q3yQlRUSgjAIAqJa+wWPO2pOmjlYeVcvbS8278vDx0Z8dIjejeUHVr+JmcEOWNMgIAqJKKiu1atDNDH/50SLvTLy2Y5mm16NbocD3Uo6GahfLvvKugjAAAqjTDMLTqwGlN/umQ1h0+49jf65raerhHI10bFcRtwU6OMgIAcBrbUs/r44TD+n5nun5eqkTRETY9dF0j3dQqVB7cFuyUKCMAAKdz5HSu/r36sOZsTlN+0aW1SiKD/DSiW0MNjo3gDhwnQxkBADit0zn5mrruqL5cd0TnLhRKkmr4e+neTvV1b+f6CgnwNTkhSoIyAgBwehcLijU3MVX/Xp2so2cu3YHj7WHVwJhwjejeUE3rBJicEL+HMgIAcBnFdkNLdmVoyqrD2pJy3rG/R9Pa+r/uDdW1cTCTXasgyggAwCUlHj2rKQnJ+mF3huNpwc1CA/RAtyjd1jZcPp4e5gaEA2UEAODSjp7J1Wdrjuirzam6UHDpGTi1qvvo3k71dU+negqu7mNyQlBGAABuIfNCoWZtStHna48oPTNPkuTtadWf2tbV/d0asIiaiSgjAAC3Ulhs16Id6fp0dbK2pWU69ndrXEv3d22gXteEyMp6JZWKMgIAcEuGYWhLyjl9sjpZi3dmOBZRaxDsr+FdGmhQbKSq+7BeSWWgjAAA3F7auQv6Yu0RzdqUquy8IklSgI+nBsdG6r4u9VU/uJrJCV0bZQQAgJ/l5hcpfkuaPlt7RIdP5UqSLBap9zUhGt61gbo1rsWtwRWAMgIAwP+w2w0lHDilz9ce0U/7Tjn2N6pdTfd1aaDb20VwCaccUUYAAPgdh0/laOq6o5qzOVW5P98aHODjqTvaRyiuc301rF3d5ITOjzICAEAJZOcVal5imqauO6rDp3Md+69rWltxneqrV7MQnhpcRpQRAABKwW43tOrgaU1de0TL9510rO4aUdNP93Sqr6GxkapZzdvckE6GMgIAQBmlnLmgaRuOavamVGVevPTUYG9Pqwa0CVdc5/qKjqxhbkAnQRkBAOAqXSwo1jfbjmvq+iPaeSzLsb9NhE33dKqvW6PD5evFs3CuhDICAEA5MQxDW1PPa9q6o/p2e7oKiu2SJJuflwa1j9Dd19ZjwutvoIwAAFABzuTka05imqatP6q0cxcd+7s2DtY919bXDS3qyMvDamLCqoMyAgBABSq2G0rYf0rT1h+9bMJrSICP7uwQqaEd66luDT9zQ5qMMgIAQCVJO3dBMzemaPamVJ3OKZAkWS1Sr2tCdHeneurR1D1vD6aMAABQyQqK7FqyO0PT16do3eEzjv3hNl8N7VBPQztEKtTma2LCykUZAQDARIdO5WjGhhTN25Km8xcu3R5stUi9m9XR3dfW03VNa7v8tyWUEQAAqoC8wmIt3pmhGRtStPHIWcf+cJuvhnSI1JDYSIW76NwSyggAAFXMwZPZmrEhVfFbL/+2pOc1IbqzQ6R6NQtxqTtxKCMAAFRReYXF+mFXhmZuTNH6w//5tqR2gI8Gt4/Q0A6Rqh9czcSE5YMyAgCAEzh8KkezN6Vq3pY0x504ktS5YbCGdojUTa1CnXaVV8oIAABOpKDIruV7T2jmxlQlHDjlWLck0NdTt7Wtq6EdItWqrs3ckKVEGQEAwEmlnbuguYlpmrM5TcfO/2eV1xZhgRraIVK3tQ1XDf+q/wRhyggAAE7Obje05tBpzd6UqiW7TjieiePtYVXflnU0JDZSXRvXqrK3CFNGAABwIedyC/R10jHN3pymPen/eYJwmM1Xd7SL0KD2EWpQq2pNeqWMAADgonYey9RXm1P1ddJxZV4sdOzv2CBIg2Ij1K91mKr7eJqY8BLKCAAALi6vsFjL9pzQnM1pWnXglOw/f6L7eXno5tahGtQ+Qp2igmU16TIOZQQAADeSkZmneVvSNC8xTYdP5zr2163hp9vb1dXt7SIUVcmXcSgjAAC4IcMwtDX1vOYmpumbbceVnVfkeK19/Zq6vV1d3dImXDY/rwrPQhkBAMDN5RUWa8nuE5qXePllHG9Pq/q0qKM72tVV9ya1K2wJesoIAABwOJGVpwVbj2neljTtP5Hj2F+rurduja6rYR0j1aROQLm+Z0k/v82fagsAACpcnUBfPdyjkR66rqF2Hc9S/JZj+jrpmE7nFOjTNclqFFKt3MtISVFGAABwIxaLRa3q2tSqrk0T+jXTqgOnNH/rcd3SOty0TKW+SJSQkKABAwYoPDxcFotFCxYs+N3j4+Pj1adPH9WuXVuBgYHq3Lmzfvjhh7LmBQAA5cTLw6rezerovWExsvlX/ITWKyl1GcnNzVV0dLQmTZpUouMTEhLUp08fLVq0SImJierVq5cGDBigrVu3ljosAABwPVc1gdVisWj+/PkaOHBgqc5r2bKlhg4dqueff75ExzOBFQAA51NlJ7Da7XZlZ2crKCjoisfk5+crPz/f8XNWVtYVjwUAAM6tYm4s/h1vvPGGcnNzNWTIkCseM3HiRNlsNscWGRlZiQkBAEBlqtQyMnPmTL344ouaPXu2QkJCrnjchAkTlJmZ6dhSU1MrMSUAAKhMlXaZZvbs2XrwwQc1Z84c3XDDDb97rI+Pj3x8fCopGQAAMFOlfDMyc+ZMDR8+XDNmzFD//v0r4y0BAICTKPU3Izk5OTp48KDj5+TkZCUlJSkoKEj16tXThAkTdOzYMU2dOlXSpSISFxend955R506dVJGRoYkyc/PTzabrZx+DQAA4KxK/c3I5s2bFRMTo5iYGEnS+PHjFRMT47hNNz09XSkpKY7jP/roIxUVFWnUqFEKCwtzbOPGjSunXwEAADgzHpQHAAAqREk/vyv91l4AAID/RhkBAACmoowAAABTVfpy8GXxy7QWloUHAMB5/PK5/UfTU52ijGRnZ0sSy8IDAOCEsrOzf3c5D6e4m8Zut+v48eMKCAiQxWIptz83KytLkZGRSk1N5S6dCsZYVy7Gu/Iw1pWHsa485TXWhmEoOztb4eHhslqvPDPEKb4ZsVqtioiIqLA/PzAwkL/YlYSxrlyMd+VhrCsPY115ymOsS7LAKRNYAQCAqSgjAADAVG5dRnx8fPTCCy/whOBKwFhXLsa78jDWlYexrjyVPdZOMYEVAAC4Lrf+ZgQAAJiPMgIAAExFGQEAAKaijAAAAFO5dRn54IMPFBUVJV9fX7Vv316rVq0yO5LTmzhxojp06KCAgACFhIRo4MCB2rdv32XHGIahF198UeHh4fLz81PPnj21a9cukxK7hokTJ8pisejRRx917GOcy9exY8d0zz33KDg4WP7+/mrbtq0SExMdrzPe5aOoqEjPPfecoqKi5Ofnp4YNG+rvf/+77Ha74xjGumwSEhI0YMAAhYeHy2KxaMGCBZe9XpJxzc/P15gxY1SrVi1Vq1ZNt956q9LS0q4+nOGmZs2aZXh5eRlTpkwxdu/ebYwbN86oVq2acfToUbOjObUbb7zR+Oyzz4ydO3caSUlJRv/+/Y169eoZOTk5jmNeffVVIyAgwJg3b56xY8cOY+jQoUZYWJiRlZVlYnLntXHjRqNBgwZGmzZtjHHjxjn2M87l5+zZs0b9+vWN4cOHGxs2bDCSk5ONZcuWGQcPHnQcw3iXj5dfftkIDg42vv32WyM5OdmYM2eOUb16dePtt992HMNYl82iRYuMZ5991pg3b54hyZg/f/5lr5dkXEeOHGnUrVvXWLp0qbFlyxajV69eRnR0tFFUVHRV2dy2jHTs2NEYOXLkZfuaNWtmPP300yYlck0nT540JBkrV640DMMw7Ha7ERoaarz66quOY/Ly8gybzWZ8+OGHZsV0WtnZ2UaTJk2MpUuXGj169HCUEca5fD311FNGt27drvg6411++vfvbzzwwAOX7bv99tuNe+65xzAMxrq8/G8ZKcm4nj9/3vDy8jJmzZrlOObYsWOG1Wo1Fi9efFV53PIyTUFBgRITE9W3b9/L9vft21dr1641KZVryszMlCQFBQVJkpKTk5WRkXHZ2Pv4+KhHjx6MfRmMGjVK/fv31w033HDZfsa5fC1cuFCxsbEaPHiwQkJCFBMToylTpjheZ7zLT7du3fTjjz9q//79kqRt27Zp9erV6tevnyTGuqKUZFwTExNVWFh42THh4eFq1arVVY+9Uzwor7ydPn1axcXFqlOnzmX769Spo4yMDJNSuR7DMDR+/Hh169ZNrVq1kiTH+P7W2B89erTSMzqzWbNmacuWLdq0adOvXmOcy9fhw4c1efJkjR8/Xs8884w2btyosWPHysfHR3FxcYx3OXrqqaeUmZmpZs2aycPDQ8XFxfrHP/6hYcOGSeLvdkUpybhmZGTI29tbNWvW/NUxV/vZ6ZZl5BcWi+Wynw3D+NU+lN3o0aO1fft2rV69+levMfZXJzU1VePGjdOSJUvk6+t7xeMY5/Jht9sVGxurV155RZIUExOjXbt2afLkyYqLi3Mcx3hfvdmzZ2vatGmaMWOGWrZsqaSkJD366KMKDw/Xfffd5ziOsa4YZRnX8hh7t7xMU6tWLXl4ePyqyZ08efJXrRBlM2bMGC1cuFArVqxQRESEY39oaKgkMfZXKTExUSdPnlT79u3l6ekpT09PrVy5Uu+++648PT0dY8k4l4+wsDC1aNHisn3NmzdXSkqKJP5el6cnnnhCTz/9tO688061bt1a9957rx577DFNnDhREmNdUUoyrqGhoSooKNC5c+eueExZuWUZ8fb2Vvv27bV06dLL9i9dulRdunQxKZVrMAxDo0ePVnx8vJYvX66oqKjLXo+KilJoaOhlY19QUKCVK1cy9qVw/fXXa8eOHUpKSnJssbGxuvvuu5WUlKSGDRsyzuWoa9euv7pFff/+/apfv74k/l6XpwsXLshqvfyjycPDw3FrL2NdMUoyru3bt5eXl9dlx6Snp2vnzp1XP/ZXNf3Vif1ya+8nn3xi7N6923j00UeNatWqGUeOHDE7mlN75JFHDJvNZvz0009Genq6Y7tw4YLjmFdffdWw2WxGfHy8sWPHDmPYsGHcllcO/vtuGsNgnMvTxo0bDU9PT+Mf//iHceDAAWP69OmGv7+/MW3aNMcxjHf5uO+++4y6des6bu2Nj483atWqZTz55JOOYxjrssnOzja2bt1qbN261ZBkvPnmm8bWrVsdS1qUZFxHjhxpREREGMuWLTO2bNli9O7dm1t7r9b7779v1K9f3/D29jbatWvnuP0UZSfpN7fPPvvMcYzdbjdeeOEFIzQ01PDx8TGuu+46Y8eOHeaFdhH/W0YY5/L1zTffGK1atTJ8fHyMZs2aGR9//PFlrzPe5SMrK8sYN26cUa9ePcPX19do2LCh8eyzzxr5+fmOYxjrslmxYsVv/vt83333GYZRsnG9ePGiMXr0aCMoKMjw8/MzbrnlFiMlJeWqs1kMwzCu7rsVAACAsnPLOSMAAKDqoIwAAABTUUYAAICpKCMAAMBUlBEAAGAqyggAADAVZQQAAJiKMgIAAExFGQEAAKaijAAAAFNRRgAAgKkoIwAAwFT/D7KDFk12+60XAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "t=np.arange(0,iterations)\n",
    "plt.plot(t,Cost_list)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "70cf86ef-4aa8-4289-8869-6d3d3e8f348f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def accuracy(inp, labels, parameters):\n",
    "    forward_cache = forward_pro(inp, parameters)\n",
    "    a_out = forward_cache['a2']   # containes propabilities with shape(10, 1)\n",
    "    \n",
    "    a_out = np.argmax(a_out, 0)  # 0 represents row wise \n",
    "    \n",
    "    labels = np.argmax(labels, 0)\n",
    "    \n",
    "    acc = np.mean(a_out == labels)*100\n",
    "    \n",
    "    return acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b4d9f0fa-ae1f-4c3d-97c4-59cd3a6ae231",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of Train Dataset 79.10000000000001 %\n",
      "Accuracy of Test Dataset 75.43 %\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy of Train Dataset\", accuracy(X_train, Y_train, Parameters), \"%\")\n",
    "print(\"Accuracy of Test Dataset\", round(accuracy(X_test, Y_test, Parameters), 2), \"%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "43fdb249-bdfc-45a0-87ee-2dbb0e9fe983",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Our model says it is : 3\n"
     ]
    }
   ],
   "source": [
    "idx = int(random.randrange(0,X_test.shape[1]))\n",
    "plt.imshow(X_test[:, idx].reshape((28,28)),cmap='gray')\n",
    "plt.show()\n",
    "\n",
    "cache = forward_pro(X_test[:, idx].reshape(X_test[:, idx].shape[0], 1), Parameters)\n",
    "a_pred = cache['a2']  \n",
    "a_pred = np.argmax(a_pred, 0)\n",
    "\n",
    "print(\"Our model says it is :\", a_pred[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e366432d-076e-44ab-a23f-ef78c752225e",
   "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}