quantum_classifier_1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "quantum_classifier_1.ipynb",
      "provenance": [],
      "toc_visible": true
    },
    "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": "3.7.6"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zzBv5XkTGur2"
      },
      "source": [
        "# Installations and Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9-xShbR18N_G"
      },
      "source": [
        "!pip install pennylane --upgrade\n",
        "\n",
        "from IPython.display import clear_output\n",
        "clear_output(wait=False)"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lp2cLbPS8RhG"
      },
      "source": [
        "import pennylane as qml\n",
        "from pennylane import numpy as np\n",
        "from pennylane.templates import RandomLayers\n",
        "\n",
        "\n",
        "#import tensorflow as tf\n",
        "#from tensorflow import keras\n",
        "\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "sns.set_theme()"
      ],
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w38G4jmvGyzk"
      },
      "source": [
        "# Importing data files"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LHAjYB6-Mhy9"
      },
      "source": [
        "### npy data files created by the quanvolutional notebook has been imported here. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gOJtRzhk7pXk"
      },
      "source": [
        "x_train_final = np.load(\"x_train_final.npy\")\n",
        "y_train = np.load(\"y_train.npy\")\n",
        "\n",
        "x_test_final = np.load(\"x_test_final.npy\")\n",
        "y_test = np.load(\"y_test.npy\")"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3z0DcNmj8jWB",
        "outputId": "ac50602a-fb9c-401f-f08e-1e6e5ead1d85"
      },
      "source": [
        "x_train_final.shape, y_train.shape, x_test_final.shape, y_test.shape"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((250, 256), (250,), (65, 256), (65,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oy37edV6Mu-G"
      },
      "source": [
        "#### train and test data has been concatenated for further size reduction as 256 features are hard to be encoded on a quantum circuit. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ebzr82i2LoGO",
        "outputId": "001c51ab-5872-40ee-f8a5-3ecf1fe39ebf"
      },
      "source": [
        "X = np.concatenate((x_train_final, x_test_final))\n",
        "Y = np.concatenate((y_train, y_test))\n",
        "print(X.shape)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(315, 256)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "T3C_OXBWNF90"
      },
      "source": [
        "#### from 256 features, 11 feautes has been selected that will be used further.  "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EbIaiDf2LoGP",
        "outputId": "ed02fbfa-5b91-4223-94a4-cf1cf32b4f1c"
      },
      "source": [
        "from sklearn.svm import LinearSVC\n",
        "from sklearn.feature_selection import SelectFromModel\n",
        "lsvc = LinearSVC(C=0.01, penalty=\"l1\", dual=False).fit(X, Y)\n",
        "model = SelectFromModel(lsvc, prefit=True)\n",
        "X = model.transform(X)\n",
        "print(X.shape)\n",
        "\n"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(315, 11)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OV0PX3SMNSif"
      },
      "source": [
        "#### Train and test data has been seperated again. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hU8stCyPLoGQ"
      },
      "source": [
        "x_train_final = (X[:250])\n",
        "x_test_final = (X[250:])"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 284
        },
        "id": "VyDjP3oALoGS",
        "outputId": "d0fbf1c8-5ef4-4484-f268-107343a848f9"
      },
      "source": [
        "from matplotlib.lines import Line2D\n",
        "from matplotlib.patches import Patch\n",
        "\n",
        "colours = [\"#ec6f86\", \"#4573e7\", \"#ad61ed\"]\n",
        "\n",
        "\n",
        "def plot_points(x_train, y_train, x_test, y_test):\n",
        "    c_train = []\n",
        "    c_test = []\n",
        "\n",
        "    for y in y_train:\n",
        "        c_train.append(colours[y])\n",
        "\n",
        "    for y in y_test:\n",
        "        c_test.append(colours[y])\n",
        "\n",
        "    plt.scatter(x_train[:, 0], x_train[:, 1], c=c_train)\n",
        "    plt.scatter(x_test[:, 0], x_test[:, 1], c=c_test, marker=\"x\")\n",
        "\n",
        "    plt.xlabel(\"Feature 1\", fontsize=16)\n",
        "    plt.ylabel(\"Feature 2\", fontsize=16)\n",
        "\n",
        "    ax = plt.gca()\n",
        "    ax.set_aspect(1)\n",
        "\n",
        "    c_transparent = \"#00000000\"\n",
        "\n",
        "    custom_lines = [\n",
        "        Patch(facecolor=colours[0], edgecolor=c_transparent, label=\"Class 0\"),\n",
        "        Patch(facecolor=colours[1], edgecolor=c_transparent, label=\"Class 1\"),\n",
        "        Patch(facecolor=colours[2], edgecolor=c_transparent, label=\"Class 2\"),\n",
        "        Line2D([0], [0], marker=\"o\", color=c_transparent, label=\"Train\",\n",
        "               markerfacecolor=\"black\", markersize=10),\n",
        "        Line2D([0], [0], marker=\"x\", color=c_transparent, label=\"Test\",\n",
        "               markerfacecolor=\"black\", markersize=10),\n",
        "    ]\n",
        "\n",
        "    ax.legend(handles=custom_lines, bbox_to_anchor=(1.0, 0.75))\n",
        "\n",
        "\n",
        "plot_points(x_train_final, y_train, x_test_final, y_test)\n"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6xrkEWrTNp-c"
      },
      "source": [
        "# Classifier Model description\n",
        "\n",
        "### In the dataset we have three classses - Covid19, Viral Penumonia and Normal person. \n",
        "\n",
        "### Our approach is to divide the classfier in two different models. First model will classify between the classes 'Normal Person' and 'Covid19/Viral Pneumonia'. Second model will classify between the classes 'Covid' and 'Viral Penumonia' \n",
        "\n",
        "### First model here is denoted by Model-1 and second one by Model-2. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bbj3YRynG3i0"
      },
      "source": [
        "# Required functions for classifier model-1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JvoQMcWF8r4Z"
      },
      "source": [
        "def get_angles(x):\n",
        "\n",
        "    beta0 = 2 * np.arcsin(np.sqrt(x[1] ** 2) / np.sqrt(x[0] ** 2 + x[1] ** 2 + 1e-12))\n",
        "    beta1 = -2 * np.arcsin(np.sqrt(x[1] ** 2) / np.sqrt(x[2] ** 2 + x[1] ** 2 + 1e-12))\n",
        "    beta2 = 2 * np.arcsin(np.sqrt(x[2] ** 2 + x[1] ** 2)/np.sqrt(x[0] ** 2 + x[1] ** 2 + x[2] ** 2 ))\n",
        "\n",
        "    return np.array([beta2, -beta1 / 2, beta1 / 2, -beta0 / 2, beta0 / 2])\n",
        "\n",
        "    "
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WmRvuHrfG96o"
      },
      "source": [
        "def statepreparation_1(a):\n",
        "\n",
        "        qml.Hadamard(wires=0)\n",
        "        qml.Hadamard(wires=1)\n",
        "        \n",
        "        qml.RZ(a[0], wires=0)\n",
        "        qml.RZ(a[1], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "        qml.RZ(a[0]*a[1], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "\n",
        "        qml.Hadamard(wires=0)\n",
        "        qml.Hadamard(wires=1)\n",
        "        \n",
        "        qml.RZ(a[1], wires=0)\n",
        "        qml.RZ(a[2], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "        qml.RZ(a[1]*a[2], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "        \n",
        "\n",
        "\n",
        "        qml.RY(a[0], wires=0)\n",
        "\n",
        "        qml.CNOT(wires=[0, 1])\n",
        "        qml.RY(a[1], wires=1)\n",
        "        qml.CNOT(wires=[0, 1])\n",
        "        qml.RY(a[2], wires=1)\n",
        "\n",
        "        for i in range(3,10):\n",
        "            qml.PauliX(wires=0)\n",
        "            qml.CNOT(wires=[0, 1])\n",
        "            qml.RY(a[i], wires=1)\n",
        "            qml.CNOT(wires=[0, 1])\n",
        "            qml.RY(a[i+1], wires=1)\n",
        "            qml.PauliX(wires=0)\n",
        "        "
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L5MOaNj_HKa2"
      },
      "source": [
        "def layer_1(W):\n",
        "    qml.Rot(W[0, 0], W[0, 1], W[0, 2], wires=0)\n",
        "    qml.Rot(W[1, 0], W[1, 1], W[1, 2], wires=1)\n",
        "    # qml.Rot(W[2, 0], W[2, 1], W[2, 2], wires=2)\n",
        "    \n",
        "    qml.CNOT(wires=[0, 1])\n",
        "    # qml.CNOT(wires=[0, 2])\n",
        "    # qml.CNOT(wires=[1, 2])"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CjT9YycqHMvK"
      },
      "source": [
        "dev_1 = qml.device(\"default.qubit\", wires=2)\n",
        "@qml.qnode(dev_1)\n",
        "def circuit_1(weights, x=None):\n",
        "    # Feature mapping\n",
        "    # angle = get_angles(x)\n",
        "    statepreparation_1(x)\n",
        "    # variational classifier\n",
        "    for w in weights:\n",
        "        layer_1(w)\n",
        "    \n",
        "    return qml.expval(qml.PauliZ(1))"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pE4EPKw1HUH3"
      },
      "source": [
        "def classifier_training_1(params, x=None, y=None):\n",
        "    weights = params[0]\n",
        "    bias = params[1]\n",
        "\n",
        "    out_probs = circuit_1(weights, x=x) + bias\n",
        "    return (out_probs-y)**2\n"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "26WdKy7PHxgQ"
      },
      "source": [
        "def classifier_prediction_1(params, x=None):\n",
        "    weights = params[0]\n",
        "    bias = params[1]\n",
        "\n",
        "    out_probs = circuit_1(weights, x=x) + bias\n",
        "\n",
        "    \n",
        "    if(out_probs>0):\n",
        "        return 1\n",
        "    else:\n",
        "        return -1"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x1jtRUU2H3rF"
      },
      "source": [
        "def circuit_output_test(params, x=None):\n",
        "    weights = params[0]\n",
        "    bias = params[1]\n",
        "\n",
        "    out_probs = circuit(weights, x=x) + bias\n",
        "\n",
        "    return out_probs\n"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h8zNvXOuH7F4"
      },
      "source": [
        " def cost_1(params, X, Y):\n",
        "\n",
        "    y_pred = np.array([classifier_training_1(params, x=X[i], y=Y[i]) for i in range(len(Y))])\n",
        "\n",
        "    cost = np.sum(y_pred) / len(Y)\n",
        "    return cost"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PnwvdZSDH-c8"
      },
      "source": [
        "def accuracy_1(params, x_train, y_train, iter):\n",
        "\n",
        "    y_pred_train = np.array([classifier_prediction_1(params, x=x) for x in x_train])\n",
        "    acc_train = np.sum(y_pred_train==y_train) / len(y_train)\n",
        "\n",
        "    print(\"Iter=> {}   train_cost=> {}   train_acc=> {} \".format(iter+1, cost_1(params, x_train, y_train), acc_train))\n",
        "\n",
        "    return acc_train\n"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xaKqVa18IHz6"
      },
      "source": [
        "# Required functions for classifier model-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tMNtf9DcIDmO"
      },
      "source": [
        "def statepreparation_2(a):\n",
        "\n",
        "        qml.Hadamard(wires=0)\n",
        "        qml.Hadamard(wires=1)\n",
        "        \n",
        "        qml.RZ(a[0], wires=0)\n",
        "        qml.RZ(a[1], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "        qml.RZ(a[0]*a[1], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "\n",
        "        qml.Hadamard(wires=0)\n",
        "        qml.Hadamard(wires=1)\n",
        "        \n",
        "        qml.RZ(a[1], wires=0)\n",
        "        qml.RZ(a[2], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "        qml.RZ(a[1]*a[2], wires=1)\n",
        "        qml.CNOT(wires=[0,1])\n",
        "        \n",
        "\n",
        "\n",
        "        qml.RY(a[0], wires=0)\n",
        "\n",
        "        qml.CNOT(wires=[0, 1])\n",
        "        qml.RY(a[1], wires=1)\n",
        "        qml.CNOT(wires=[0, 1])\n",
        "        qml.RY(a[2], wires=1)\n",
        "\n",
        "        for i in range(3,10):\n",
        "            qml.PauliX(wires=0)\n",
        "            qml.CNOT(wires=[0, 1])\n",
        "            qml.RY(a[i], wires=1)\n",
        "            qml.CNOT(wires=[0, 1])\n",
        "            qml.RY(a[i+1], wires=1)\n",
        "            qml.PauliX(wires=0)"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HAXZ-a1aIQqG"
      },
      "source": [
        "def layer_2(W):\n",
        "    qml.Rot(W[0, 0], W[0, 1], W[0, 2], wires=0)\n",
        "    qml.Rot(W[1, 0], W[1, 1], W[1, 2], wires=1)\n",
        "    # qml.Rot(W[2, 0], W[2, 1], W[2, 2], wires=2)\n",
        "    \n",
        "    qml.CNOT(wires=[0, 1])\n",
        "    # qml.CNOT(wires=[0, 2])\n",
        "    # qml.CNOT(wires=[1, 2])"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9EZJ7OJcIT62"
      },
      "source": [
        "dev_2 = qml.device(\"default.qubit\", wires=2)\n",
        "@qml.qnode(dev_2)\n",
        "def circuit_2(weights, x=None):\n",
        "    # Feature mapping\n",
        "    # angle = get_angles(x)\n",
        "    statepreparation_2(x)\n",
        "    \n",
        "    # variational classifier\n",
        "    for w in weights:\n",
        "        layer_2(w)\n",
        "    \n",
        "    return qml.expval(qml.PauliZ(1))"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iumVNlMWIcDg"
      },
      "source": [
        "def classifier_training_2(params, x=None, y=None):\n",
        "    weights = params[0]\n",
        "    bias = params[1]\n",
        "\n",
        "    out_probs = circuit_2(weights, x=x) + bias\n",
        "    return (out_probs-y)**2"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aC47T9L6Ih2V"
      },
      "source": [
        "def classifier_prediction_2(params, x=None):\n",
        "    weights = params[0]\n",
        "    bias = params[1]\n",
        "\n",
        "    out_probs = circuit_2(weights, x=x) + bias\n",
        "\n",
        "    \n",
        "    if(out_probs>0):\n",
        "        return 1\n",
        "    else:\n",
        "        return -1"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lix_uPNYIlJT"
      },
      "source": [
        "def cost_2(params, X, Y):\n",
        "\n",
        "    y_pred = np.array([classifier_training_2(params, x=X[i], y=Y[i]) for i in range(len(Y))])\n",
        "\n",
        "    cost = np.sum(y_pred) / len(Y)\n",
        "    return cost\n"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yZAVHcVOIoXE"
      },
      "source": [
        "def accuracy_2(params, x_train, y_train, iter):\n",
        "\n",
        "    y_pred_train = np.array([classifier_prediction_2(params, x=x) for x in x_train])\n",
        "    acc_train = np.sum(y_pred_train==y_train) / len(y_train)\n",
        "\n",
        "    print(\"Iter=> {}   train_cost=> {}   train_acc=> {} \".format(iter+1, cost_2(params, x_train, y_train), acc_train))\n",
        "\n",
        "    return acc_train\n",
        "    "
      ],
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3-jPuIBwItKu"
      },
      "source": [
        "# Model-1"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3BFJgYEPKMlX"
      },
      "source": [
        "## Data for model-1\n",
        "\n",
        "0->covid, 1->Normal, 2->Penumonia"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7QKas5BHO7MP"
      },
      "source": [
        "#### Here the labels are modified accordinglly for Model-1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DC-vze_qIrlM"
      },
      "source": [
        "x_train_1 = np.copy(x_train_final)\n",
        "y_train_1 = []\n",
        "\n",
        "for i in range(len(y_train)):\n",
        "    if(y_train[i]==0 or y_train[i]==2):  ##Covid or pneumonia\n",
        "        y_train_1.append(-1)\n",
        "    elif(y_train[i]==1):  ##normal person\n",
        "        y_train_1.append(1)\n",
        "\n",
        "y_train_1 = np.array(y_train_1)"
      ],
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xhZkWLIDLjiY"
      },
      "source": [
        "## Binary Classifier Model-1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YylpUnE0K7Y5",
        "outputId": "537a529a-48ea-4cd0-98b2-0793c8db202e"
      },
      "source": [
        "params_1 = (0.01 * np.random.randn(2, 2, 3), 0.0)\n",
        "params_1"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(tensor([[[ 0.00788197,  0.00862789, -0.01362923],\n",
              "          [ 0.03096347, -0.00686296,  0.00633331]],\n",
              " \n",
              "         [[-0.01199092, -0.01398299,  0.00185472],\n",
              "          [ 0.00721716,  0.00810696,  0.00811893]]], requires_grad=True), 0.0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nHkAyNQ3LTbW",
        "outputId": "2a905a7d-ba4c-43db-f51a-52a52db82f93"
      },
      "source": [
        "iters = 20\n",
        "optimizer_1 = qml.AdagradOptimizer(stepsize=0.5)\n",
        "\n",
        "cost_list_1 = []\n",
        "train_acc_list_1 = []\n",
        "\n",
        "for iter in range(iters):\n",
        "    params_1 = optimizer_1.step(lambda v: cost_1(v, x_train_1, y_train_1), params_1)\n",
        "\n",
        "    cost_list_1.append(cost_1(params_1, x_train_1, y_train_1))\n",
        "    train_acc_list_1.append(accuracy_1(params_1, x_train_1, y_train_1, iter))"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Iter=> 1   train_cost=> 0.9430680970436425   train_acc=> 0.636 \n",
            "Iter=> 2   train_cost=> 0.9434245396114554   train_acc=> 0.588 \n",
            "Iter=> 3   train_cost=> 0.9049820852738587   train_acc=> 0.688 \n",
            "Iter=> 4   train_cost=> 0.8701317914427136   train_acc=> 0.664 \n",
            "Iter=> 5   train_cost=> 0.8235834174948404   train_acc=> 0.72 \n",
            "Iter=> 6   train_cost=> 0.8041318123394667   train_acc=> 0.712 \n",
            "Iter=> 7   train_cost=> 0.7895082700859484   train_acc=> 0.712 \n",
            "Iter=> 8   train_cost=> 0.7856300343953608   train_acc=> 0.716 \n",
            "Iter=> 9   train_cost=> 0.7992519245856506   train_acc=> 0.696 \n",
            "Iter=> 10   train_cost=> 0.794766268933636   train_acc=> 0.692 \n",
            "Iter=> 11   train_cost=> 0.7820467651355429   train_acc=> 0.7 \n",
            "Iter=> 12   train_cost=> 0.7768421984495323   train_acc=> 0.692 \n",
            "Iter=> 13   train_cost=> 0.7706342494969206   train_acc=> 0.704 \n",
            "Iter=> 14   train_cost=> 0.769017766266119   train_acc=> 0.692 \n",
            "Iter=> 15   train_cost=> 0.7669294220914374   train_acc=> 0.708 \n",
            "Iter=> 16   train_cost=> 0.7667130993171604   train_acc=> 0.7 \n",
            "Iter=> 17   train_cost=> 0.7660947881445909   train_acc=> 0.704 \n",
            "Iter=> 18   train_cost=> 0.767394522447745   train_acc=> 0.708 \n",
            "Iter=> 19   train_cost=> 0.7666808888269254   train_acc=> 0.7 \n",
            "Iter=> 20   train_cost=> 0.7680832338923542   train_acc=> 0.704 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_0KN2j9lWI_R"
      },
      "source": [
        "## Cost Plot of Model-1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        },
        "id": "TnQU5gmrVimZ",
        "outputId": "24808797-113b-4e1d-d5ba-c0a852059ffd"
      },
      "source": [
        "plt.plot(cost_list_1)"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f90de0cbed0>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6mQU4saMWMxy"
      },
      "source": [
        "## Training Accuracy Plot of Model-1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "MPXF2BLwVoIZ",
        "outputId": "1921f230-0663-41be-9018-0bbecdcba47b"
      },
      "source": [
        "plt.plot(train_acc_list_1)"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f90deb12a90>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EuG6oxpIR95T"
      },
      "source": [
        "# Model-2"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "up0q4ehrR_9r"
      },
      "source": [
        "## Data for model-2\n",
        "0->covid, 1->Normal, 2->Penumonia"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kqB8sMaiPHKj"
      },
      "source": [
        "#### Here the labels are modified accordingly for Model-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MG-55rGiSGaI"
      },
      "source": [
        "x_train_2 = []\n",
        "y_train_2 = []\n",
        "\n",
        "for i in range(len(y_train)):\n",
        "    if(y_train[i]==0):   ## covid\n",
        "        x_train_2.append(x_train_final[i])\n",
        "        y_train_2.append(1)\n",
        "    elif(y_train[i]==2):  ## pneumonia\n",
        "        x_train_2.append(x_train_final[i])\n",
        "        y_train_2.append(-1)\n",
        "\n",
        "\n",
        "x_train_2 = np.array(x_train_2)\n",
        "y_train_2 = np.array(y_train_2)"
      ],
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d-DYdwjIS4_T"
      },
      "source": [
        "## Binary Classifier Model-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "osCtRvneS3FM",
        "outputId": "1efd789b-fb89-4633-947d-bd0065212012"
      },
      "source": [
        "params_2 = (0.01 * np.random.randn(2, 2, 3), 0.0)\n",
        "params_2"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(tensor([[[-0.01642739, -0.00303444,  0.00879709],\n",
              "          [ 0.00396784,  0.03160051, -0.00861578]],\n",
              " \n",
              "         [[ 0.00302629,  0.00091985,  0.00371899],\n",
              "          [ 0.01023066, -0.00811128, -0.00915326]]], requires_grad=True), 0.0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uKoRU8xBTEiL",
        "outputId": "03f95d69-e7e9-4469-da64-86e06e1d561e"
      },
      "source": [
        "iters = 20\n",
        "optimizer_2 = qml.AdagradOptimizer(stepsize=0.5)\n",
        "\n",
        "cost_list_2 = []\n",
        "train_acc_list_2 = []\n",
        "\n",
        "for iter in range(iters):\n",
        "    params_2 = optimizer_2.step(lambda v: cost_2(v, x_train_2, y_train_2), params_2)\n",
        "\n",
        "    cost_list_2.append(cost_2(params_2, x_train_2, y_train_2))\n",
        "    train_acc_list_2.append(accuracy_2(params_2, x_train_2, y_train_2, iter))"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Iter=> 1   train_cost=> 1.347974932146223   train_acc=> 0.4444444444444444 \n",
            "Iter=> 2   train_cost=> 0.9856259090057776   train_acc=> 0.6388888888888888 \n",
            "Iter=> 3   train_cost=> 0.9925207251175971   train_acc=> 0.6222222222222222 \n",
            "Iter=> 4   train_cost=> 1.004129402328754   train_acc=> 0.6166666666666667 \n",
            "Iter=> 5   train_cost=> 0.8884518646533924   train_acc=> 0.6611111111111111 \n",
            "Iter=> 6   train_cost=> 0.8091313055032515   train_acc=> 0.7 \n",
            "Iter=> 7   train_cost=> 0.7829457458090165   train_acc=> 0.7111111111111111 \n",
            "Iter=> 8   train_cost=> 0.7681887981986943   train_acc=> 0.7166666666666667 \n",
            "Iter=> 9   train_cost=> 0.7599551749217474   train_acc=> 0.7222222222222222 \n",
            "Iter=> 10   train_cost=> 0.753346260706608   train_acc=> 0.7277777777777777 \n",
            "Iter=> 11   train_cost=> 0.7492179270476366   train_acc=> 0.7388888888888889 \n",
            "Iter=> 12   train_cost=> 0.7460577481431032   train_acc=> 0.7388888888888889 \n",
            "Iter=> 13   train_cost=> 0.7437081027338572   train_acc=> 0.7444444444444445 \n",
            "Iter=> 14   train_cost=> 0.7418335293994769   train_acc=> 0.7444444444444445 \n",
            "Iter=> 15   train_cost=> 0.7403042064384163   train_acc=> 0.75 \n",
            "Iter=> 16   train_cost=> 0.7390145390398503   train_acc=> 0.75 \n",
            "Iter=> 17   train_cost=> 0.7379022344378474   train_acc=> 0.7555555555555555 \n",
            "Iter=> 18   train_cost=> 0.7369217587371343   train_acc=> 0.7555555555555555 \n",
            "Iter=> 19   train_cost=> 0.7360424737515406   train_acc=> 0.75 \n",
            "Iter=> 20   train_cost=> 0.7352414799044594   train_acc=> 0.75 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RGjV2w46Wbct"
      },
      "source": [
        "## Cost Plot of Model-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "R_d6_C8AWdv1",
        "outputId": "bd14df2f-5b7d-4a59-fee0-c632af6956ae"
      },
      "source": [
        "plt.plot(cost_list_2)"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f90defe4150>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UENuiv_wWgGX"
      },
      "source": [
        "## Training Accuracy Plot of Model-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "vwcCN-_6WlC4",
        "outputId": "5912aeac-b153-4bca-feaa-86e7c22d31d6"
      },
      "source": [
        "plt.plot(train_acc_list_2)"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f90ddfd8d50>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wKuT5KBeTiQk"
      },
      "source": [
        "# Prediction"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gqLx2cuKPcVW"
      },
      "source": [
        "### While Predicting, we first give input to the Model-1. If it predicts as Normal person, then it is the final prediction assigned to the input. If not, then we give the same input to Model-2 and it finally predicts whether the chest x-ray is Covid10 patient or Viral Pneumonia patient. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Y1yxaM9Tao4"
      },
      "source": [
        "y_pred = []\n",
        "\n",
        "for i in range(len(x_test_final)):\n",
        "    tmp = classifier_prediction_1(params_1, x=x_test_final[i])\n",
        "    if (tmp == 1):  ## Normal person\n",
        "        y_pred.append(1)\n",
        "    else:   ## Covid or Pneumonia\n",
        "        tmp = classifier_prediction_2(params_2, x=x_test_final[i])\n",
        "        if(tmp == 1):  ## covid\n",
        "            y_pred.append(0)\n",
        "        elif(tmp == -1): ## Pneumonia\n",
        "            y_pred.append(2)\n",
        "\n",
        "y_pred = np.array(y_pred)"
      ],
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "r1trY_8EUuaM",
        "outputId": "e9a245f3-93e8-4cc3-d7e3-1fcc8d9a08f5"
      },
      "source": [
        "print(\"Final_Test_Accuracy => \", np.sum(y_pred==y_test)/len(y_test))"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Final_Test_Accuracy =>  0.49230769230769234\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tQzw0oEbQIAk"
      },
      "source": [
        "### We are hopeful to in increment in this accuracy when it will be trained on real computer with larger dataset. "
      ]
    }
  ]
}