Created using Colaboratory

XOR_NN.ipynb

@@ -0,0 +1,213 @@
+{
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
+    "colab": {
+      "name": "XOR-NN.ipynb",
+      "provenance": [],
+      "collapsed_sections": [],
+      "authorship_tag": "ABX9TyP/dE4A5rYjsAy2fBvCAVOU",
+      "include_colab_link": true
+    },
+    "kernelspec": {
+      "name": "python3",
+      "display_name": "Python 3"
+    },
+    "language_info": {
+      "name": "python"
+    }
+  },
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "view-in-github",
+        "colab_type": "text"
+      },
+      "source": [
+        "<a href=\"https://colab.research.google.com/github/akagam1/Machine-Learning/blob/main/XOR_NN.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "65402TIidgZH"
+      },
+      "outputs": [],
+      "source": [
+        "import numpy as np\n",
+        "import random\n",
+        "import matplotlib.pyplot as plt"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "x1 = []\n",
+        "x2 = []\n",
+        "y = []\n",
+        "A2 = 0\n",
+        "training = 100\n",
+        "for i in range(training):\n",
+        "  temp1 = random.randint(0,1)\n",
+        "  temp2 = random.randint(0,1)\n",
+        "  x1.append(temp1)\n",
+        "  x2.append(temp2)\n",
+        "\n",
+        "for i in range(training):\n",
+        "  if x1[i] != x2[i]:\n",
+        "    y.append(1)\n",
+        "  else:\n",
+        "    y.append(0)\n",
+        "\n",
+        "x = np.array([x1,x2])\n",
+        "y = np.array(y)\n",
+        "print(x.shape[1])"
+      ],
+      "metadata": {
+        "id": "9xLxBQqwd-A_",
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "outputId": "fb9606bb-18cb-4e78-8dad-dab9defc400f"
+      },
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "100\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "def sigmoid(z):\n",
+        "  return 1/(1+np.exp(-z))\n",
+        "\n",
+        "def initParas(input_size, hidden_size, output_size):\n",
+        "  w1 = np.random.randn(hidden_size, input_size)\n",
+        "  w2 = np.random.randn(output_size, hidden_size)\n",
+        "  b1 = np.zeros((hidden_size, 1))\n",
+        "  b2 = np.zeros((output_size, 1))\n",
+        "\n",
+        "  return (w1,w2, b1, b2)\n",
+        "\n",
+        "def forwardProp(x,y, parameters):\n",
+        "  global A2\n",
+        "  w1,w2,b1,b2 = parameters\n",
+        "  m = x.shape[1]\n",
+        "\n",
+        "  z1 = np.dot(w1,x) + b1\n",
+        "  a1 = sigmoid(z1)\n",
+        "  z2 = np.dot(w2,a1) + b2\n",
+        "  a2 = sigmoid(z2)\n",
+        "\n",
+        "  A2 = a2\n",
+        "  return (z1,a1,z2,a2)\n",
+        "\n",
+        "def backProp(x,y, forward,parameters):\n",
+        "  w1,w2,b1,b2 = parameters\n",
+        "  z1,a1,z2,a2= forward\n",
+        "  m = x.shape[1]\n",
+        "\n",
+        "  dz2 = (y-a2)/m\n",
+        "  dw2 = np.dot(dz2, a1.T)\n",
+        "  db2 = np.sum(dz2, axis =1, keepdims = True) \n",
+        "\n",
+        "  da1 = np.dot(w2.T, dz2)\n",
+        "  dz1 = np.multiply(da1, a1 * (1-a1))\n",
+        "  dw1 = np.dot(dz1, x.T)\n",
+        "  db1 = np.sum(dz1, axis=1, keepdims = True) \n",
+        "\n",
+        "  return (dz2, dw2, db2, dz1, dw1, db1)\n",
+        "\n",
+        "def paraUpdate(parameters, grads, alpha):\n",
+        "  w1,w2,b1,b2 = parameters\n",
+        "  w1 = w1 + alpha*grads[4]\n",
+        "  w2 = w2 + alpha*grads[1]\n",
+        "  b1 = b1 + alpha*grads[5]\n",
+        "  b2 = b2 + alpha*grads[2]\n",
+        "\n",
+        "  return (w1,w2,b1,b2)\n"
+      ],
+      "metadata": {
+        "id": "SJUT6WxxpN1X"
+      },
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "epochs = 10000\n",
+        "alpha = 0.1\n",
+        "input = 2\n",
+        "output = 1\n",
+        "hidden = 2 #can be any value\n",
+        "parameters = initParas(input,hidden,output)\n",
+        "\n",
+        "for i in range(epochs):\n",
+        "  forward = forwardProp(x,y,parameters)\n",
+        "  back = backProp(x,y, forward, parameters)\n",
+        "  parameters = paraUpdate(parameters, back, alpha)\n"
+      ],
+      "metadata": {
+        "id": "G66zH4W8zSlB"
+      },
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "x1 = []\n",
+        "x2 = []\n",
+        "ytest = []\n",
+        "tests=10\n",
+        "for i in range(tests):\n",
+        "  temp1 = random.randint(0,1)\n",
+        "  temp2 = random.randint(0,1)\n",
+        "  x1.append(temp1)\n",
+        "  x2.append(temp2)\n",
+        "\n",
+        "for i in range(tests):\n",
+        "  if x1[i] != x2[i]:\n",
+        "    ytest.append(1)\n",
+        "  else:\n",
+        "    ytest.append(0)\n",
+        "xtest = np.array([x1,x2])\n",
+        "ytest = np.array(ytest)\n",
+        "\n",
+        "correct = 0\n",
+        "temp = forwardProp(xtest,ytest,parameters)\n",
+        "for i in range(xtest.shape[1]):\n",
+        "  A2[0][i] = round(A2[0][i])\n",
+        "  if A2[0][i] == ytest[i]:\n",
+        "    correct += 1\n",
+        "print(f'Percent Accuracy: {(correct/tests)*100}%')"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "QPSubX3c7Euz",
+        "outputId": "e5fd4581-35e5-480a-f7e3-8a28fdffbcf1"
+      },
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Percent Accuracy: 100.0%\n"
+          ]
+        }
+      ]
+    }
+  ]
+}