Galaxy_Detection_using_QNN.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/MonitSharma/Learn-Quantum-Machine-Learning/blob/main/Galaxy_Detection_using_QNN.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Classifying existence of Galaxy using Quantum Machine Learning\n",
        "\n",
        "\n",
        "In the following code notebook, you'll learn how to develop a galaxy detection technique from the telescope image via Quantum Machine Learning.\n",
        "\n",
        "-----\n",
        "\n",
        "The accuracy of the model is $94\\%$ via a quantum machine learning model from NASA image. The image is further divided into a small $16 \\times 16$ image as input data. \n",
        "\n",
        "-----\n",
        "\n",
        "Then the data is encoded and trained using a quantum circuit using the parameterized quantum circuit taken from the paper [Expressibilitty and entangling capability of parameterized quantum circuit for hybrid quantum-classical algorithms](https://arxiv.org/abs/1905.10876).\n",
        "\n",
        "-----\n",
        "\n",
        "With the high expressibility of the circuit, the model is trained and tested with Cross-Entropy as the Loss function and **L-BFGS** algorithm for optimization.\n",
        "\n",
        "\n",
        "This algorithm is realized in **PyTorch** and **Qiskit** machine learning module."
      ],
      "metadata": {
        "id": "Ap7nRE9m6rXh"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Pre-Processing of Image\n",
        "\n",
        "\n",
        "Import necessary Libraries for Image pre-processing"
      ],
      "metadata": {
        "id": "4_4Ylck69OK9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from PIL import Image\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n"
      ],
      "metadata": {
        "id": "mgpX3QNh9Rwj"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Pre-Processing Step\n",
        "\n",
        "The below step takes the original image taken by the telescope which is around $3$ MB each. It reshapes the images in the $2463 \\times 2895$ dimensions and normalized, and divided into equal $16 \\times 16$ pieces. The results are named by the nymber it represents and saved in the `result` folder in grayscale."
      ],
      "metadata": {
        "id": "dLWks8DR9e3K"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cut=128\n",
        "\n",
        "img = Image.open('nasa_image.jpg').convert('L')\n",
        "plt.imshow(img)\n",
        "img = np.asarray(img.getdata()).reshape((2463,2895))\n",
        "\n",
        "for i in range(2463//cut):\n",
        "    for j in range(2895//cut):\n",
        "        cutimg = img[i*cut:(i+1)*cut]\n",
        "        cutimg = cutimg[:,list(range(j*cut,(j+1)*cut))]\n",
        "        plt.imsave('result/{} {}.jpg'.format(i,j), cutimg ,cmap='gray')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "id": "gsPbHuhm9ZXY",
        "outputId": "4f967554-8c07-4668-bdb8-d3886ef23e58"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Visualize one of the Images\n",
        "\n",
        "Here we'll visualize , one of the images from the result folder , which we will use to further train and test our data."
      ],
      "metadata": {
        "id": "gDLyahj-AYBZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "for i in range(2463//cut):\n",
        "    for j in range(2895//cut):\n",
        "        img16 = Image.open('result/{} {}.jpg'.format(i,j)).resize((16,16)).convert('L')\n",
        "        img16 = np.asarray(img16.getdata()).reshape((16,16))\n",
        "        "
      ],
      "metadata": {
        "id": "EqXEEg9BANsb"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "img16 = Image.open('result/{} {}.jpg'.format(5,1)).resize((16,16)).convert('L')\n",
        "plt.imshow(img16, cmap='gray')\n",
        "img16 = np.asarray(img16.getdata(), dtype=np.float64).reshape((16,16))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "WDYZBsRWAVBq",
        "outputId": "ca68a0cd-033e-49f7-b6bf-65f8fc4d8851"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Ideally this is done in a separate notebook, but is shown here, just for demonstration purposes."
      ],
      "metadata": {
        "id": "7ByBVPixAzLn"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Quantum Convolutional Neural Network\n",
        "\n",
        "We will now build our QCNN to detect the Galaxy in the telescope images.\n",
        "\n",
        "\n",
        "\n",
        "-------\n",
        "\n",
        "We will start by importing some helper function files we made to help us train faster with less code.\n",
        "\n",
        "\n",
        "This code will help us read the Images faster:\n",
        "\n",
        "```python\n",
        "import numpy as np\n",
        "\n",
        "#Image processing\n",
        "from PIL import Image\n",
        "\n",
        "#Calling Image from the path, and file name. \n",
        "#i is the number labeling of the data\n",
        "#some of the files does not have name\n",
        "def callImage(i,path,name):\n",
        "    x1 = Image.open(\n",
        "        path+str(name)+str(i)+'.jpg').convert('L');\n",
        "    y1 = np.asarray(x1.getdata(), dtype=np.float64).reshape((x1.size[1], x1.size[0]));\n",
        "    y_dat1 = np.asarray(y1, dtype=np.uint8)     \n",
        "    return y_dat1\n",
        "\n",
        "#Resize image into n x n pixel\n",
        "def imageResize(data,pixel):\n",
        "    image = Image.fromarray(data,'L')\n",
        "    image= image.resize((pixel, pixel))\n",
        "    image=np.asarray(image.getdata(), dtype=np.float64).reshape((image.size[1], image.size[0]))\n",
        "    image=np.asarray(image, dtype=np.uint8)    \n",
        "    return image\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "#Making MxN partition\n",
        "def imagePartition(data,M,N):\n",
        "    tiles = [data[x:x+M,y:y+N] for x in range(0,data.shape[0],M) for y in range(0,data.shape[1],N)]\n",
        "    return tiles\n",
        "\n",
        "def imageBinarize(data):\n",
        "    # specify a threshold 0-255\n",
        "    threshold = 75\n",
        "    # make all pixels < threshold black\n",
        "    bidata = 1.0 * (data > threshold)\n",
        "    return bidata\n",
        "    \n",
        "```\n",
        "\n",
        "Save this file as `ImageRead.py`\n",
        "\n",
        "\n",
        "-----\n",
        "\n",
        "Use this file to make the Quantum Neural network:\n",
        "\n",
        "```python\n",
        "import qiskit\n",
        "from qiskit import transpile, assemble\n",
        "from qiskit import QuantumCircuit, ClassicalRegister, QuantumRegister, execute\n",
        "from qiskit import BasicAer, Aer, execute\n",
        "from qiskit.quantum_info import state_fidelity\n",
        "from qiskit.visualization import *\n",
        "from qiskit.quantum_info.operators import Operator\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from qiskit.circuit.parameter import Parameter\n",
        "import torch\n",
        "from torch.autograd import Function\n",
        "from torchvision import datasets, transforms\n",
        "import torch.optim as optim\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "\n",
        "import qiskit\n",
        "from qiskit import transpile, assemble\n",
        "from qiskit.visualization import *\n",
        "\n",
        "\n",
        "nqubits=6\n",
        "\n",
        "def normlaizeData(data):\n",
        "    #Create Array of pixel value\n",
        "    testdata=data\n",
        "    arr_data=testdata.flatten()/max(testdata.flatten());\n",
        "    encoding_data= np.array([np.round(x,6) for x in arr_data]);\n",
        "    sum_const=np.sqrt(sum(encoding_data*encoding_data))\n",
        "    encoding_norm=encoding_data/sum_const\n",
        "    return encoding_norm\n",
        "\n",
        "# Choose on PQC from Hannah  Sim https://arxiv.org/pdf/1905.10876.pdf circuit 15\n",
        "\n",
        "def circuit15(qc,theta):\n",
        "    #circuit 15\n",
        "    #theta is list of the parameters\n",
        "    #theta length is (8)L\n",
        "    #L is the number of repeatation\n",
        "    nqubits=6\n",
        "    qr = QuantumRegister(nqubits)\n",
        "    qc = QuantumCircuit(qr, name='PQC')\n",
        "\n",
        "    count=0\n",
        "\n",
        "\n",
        "    for i in range(nqubits):\n",
        "        qc.ry(theta[count],i)\n",
        "        count=count+1\n",
        "    for i in range(nqubits-1):\n",
        "        qc.cx(i,i+1)\n",
        "    \n",
        "    qc.cx(0,nqubits-1)\n",
        "    for i in range(nqubits):\n",
        "        qc.ry(theta[count],i)\n",
        "        count=count+1    \n",
        "    for i in range(nqubits-1):\n",
        "        qc.cx(i+1,i)\n",
        "    qc.cx(nqubits-1,0)\n",
        "    qc.to_instruction()\n",
        "    return qc\n",
        "# Choose on PQC from Hannah  Sim https://arxiv.org/pdf/1905.10876.pdf circuit 15\n",
        "\n",
        "def encoding(qc,theta,L):\n",
        "    #circuit 15\n",
        "    #theta is list of the parameters\n",
        "    #theta length is (8)L\n",
        "    #L is the number of repeatation\n",
        "    nqubits=6\n",
        "    qr = QuantumRegister(nqubits)\n",
        "    qc = QuantumCircuit(qr, name='Embed')\n",
        "\n",
        "    count=0\n",
        "    for i in range(nqubits):\n",
        "        qc.h(i)\n",
        "        \n",
        "    for l in range(L):\n",
        "        for i in range(nqubits):\n",
        "            qc.ry(theta[count],i)\n",
        "            count=count+1\n",
        "        for i in range(nqubits-1):\n",
        "            qc.cx(i,i+1)\n",
        "        \n",
        "        qc.cx(nqubits-1,0)\n",
        "        for i in range(nqubits):\n",
        "            qc.ry(theta[count],i)\n",
        "            count=count+1    \n",
        "        for i in range(nqubits-1):\n",
        "            qc.cx(i+1,i)\n",
        "        qc.cx(0,nqubits-1)\n",
        "        \n",
        "    qc.to_instruction()\n",
        "    return qc\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# mapping the data\n",
        "# mapping is taken from https://arxiv.org/pdf/2003.09887.pdf\n",
        "def binary(x):\n",
        "    return ('0'*(6-len('{:b}'.format(x, '#010b') ))+'{:b}'.format(x, '#010b'))\n",
        "def firsttwo(x):\n",
        "    return x[:2]\n",
        "parity = lambda x: firsttwo(binary(x)).count('1') % 2   \n",
        "\n",
        "```\n",
        "\n",
        "Save this as `QNN.py`  and as you can see this file is heavily Qiskit dependent, make sure to install the latest Qiskit packages and all its necessary dependencies.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "erD6Ukl9A8-h"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "-----\n",
        "-----\n",
        "\n",
        "Import Basic python packages and the two python helper scirpts that we made to make our life easier.\n",
        "\n"
      ],
      "metadata": {
        "id": "fCnlSHdnCWkO"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PEOinaYO5cIs"
      },
      "outputs": [],
      "source": [
        "import random\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import sys\n",
        "sys.path.append('../PyFiles')   # where the Helper files are saved\n",
        "# Pull in the helper files.\n",
        "from ImageRead import *  # the helper files we created to read Images\n",
        "from QNN import *        # the helper file that creates the QNN\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "YMKRo4jl5cI7"
      },
      "outputs": [],
      "source": [
        "target_o = [1 for i in range(25)]+[0 for i in range(25)]"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Loading the dataset\n",
        "\n",
        "Load all the data files from the respective data folder and use the helper function to divide the images into smaller parts to make it easier for the QNN to read and make sense of."
      ],
      "metadata": {
        "id": "z7EuQ86NC1Bp"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "j0hhVp4p5cI_"
      },
      "outputs": [],
      "source": [
        "pathY=r'../dataset/Original/galaxy/'\n",
        "pathN=r'../dataset/Original/No-galaxy/'\n",
        "nameN=''\n",
        "nameY=''\n",
        "\n",
        "inputY=[imageResize(callImage(i+1,pathY,nameY),16) for i in range(25)]\n",
        "inputN=[imageResize(callImage(i+1,pathN,nameN),16) for i in range(25)]\n",
        "input_combine = inputY+inputN"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Randomize and Re-Shuffle the data set to make it better to train."
      ],
      "metadata": {
        "id": "u17hzIPlDNEr"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PWO4DFzv5cJK"
      },
      "outputs": [],
      "source": [
        "np.random.seed(0)\n",
        "idx=np.array([int(i) for i in range(50)]).flatten()\n",
        "\n",
        "np.random.shuffle(idx)\n",
        "\n",
        "dataInput = list(input_combine[i] for i in idx )\n",
        "dataTarget = list( imageBinarize(input_combine[i]) for i in idx )\n",
        "\n",
        "data_target_o=list( target_o[i] for i in idx )"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Visualize\n",
        "\n",
        "Display the images from the folder after all the pre-processing and ready to be feeded to the QNN."
      ],
      "metadata": {
        "id": "Cj4u8emLDUKL"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cjKji2Xa5cJM",
        "outputId": "ddc756ec-e181-48e6-fd24-b6c0e39d77aa"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1440x432 with 20 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "n_samples_show = 10\n",
        "fig, axes = plt.subplots(nrows=2, ncols=n_samples_show, figsize=(20, 6))\n",
        "\n",
        "for i in range(n_samples_show):\n",
        "\n",
        "    axes[0,i].imshow(dataInput[i], cmap='gray')\n",
        "    axes[0,i].set_xticks([])\n",
        "    axes[0,i].set_yticks([])\n",
        "    axes[1,i].imshow(dataInput[i+5], cmap='gray')\n",
        "    axes[1,i].set_xticks([])\n",
        "    axes[1,i].set_yticks([])    \n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## The Quantum Neural Network\n",
        "\n",
        "\n",
        "Import necessary files for the training."
      ],
      "metadata": {
        "id": "aBcnyyP-DeTG"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "trdX82lA5cJX"
      },
      "outputs": [],
      "source": [
        "from qiskit.circuit.parameter import Parameter\n",
        "from qiskit_machine_learning.neural_networks import CircuitQNN\n",
        "from qiskit_machine_learning.connectors import TorchConnector\n",
        "from qiskit.utils import QuantumInstance\n",
        "from torch.nn import Linear, CrossEntropyLoss, MSELoss\n",
        "from torch.optim import LBFGS, SGD,Adam \n",
        "qi = QuantumInstance(Aer.get_backend('statevector_simulator'))\n",
        "from qiskit.circuit.library import RealAmplitudes, ZZFeatureMap\n",
        "from qiskit_machine_learning.algorithms.classifiers import NeuralNetworkClassifier, VQC\n",
        "from qiskit.algorithms.optimizers import COBYLA, L_BFGS_B\n",
        "\n",
        "# Model for LBFGS\n",
        "# Combining the circuit together with CircuitQNN\n",
        "np.random.seed(3)\n",
        "\n",
        "\n",
        "nqubits=6\n",
        "num_inputs=256\n",
        "qc = QuantumCircuit(nqubits)\n",
        "\n",
        "# Encoding\n",
        "param_x=[];\n",
        "for i in range(num_inputs):\n",
        "    param_x.append(Parameter('x'+str(i)))\n",
        "for i in range(8):\n",
        "    param_x.append(np.pi/2)\n",
        "\n",
        "\n",
        "feature_map = encoding(qc,param_x,22)\n",
        "\n",
        "\n",
        "# Optimzing circuit PQC\n",
        "param_y=[];\n",
        "for i in range(nqubits*2):\n",
        "    param_y.append(Parameter('θ'+str(i)))\n",
        "\n",
        "ansatz=circuit15(qc,param_y)\n",
        "\n",
        "qc.append(feature_map, range(nqubits))\n",
        "qc.append(ansatz, range(nqubits))\n",
        "\n",
        "qnn2 = CircuitQNN(qc, input_params=feature_map.parameters, weight_params=ansatz.parameters, \n",
        "                  interpret=parity, output_shape=2, quantum_instance=qi)\n",
        "initial_weights = 0.1*(2*np.random.rand(qnn2.num_weights) - 1)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YxOGd0gn5cJ_"
      },
      "source": [
        "# Learning Rate 0.05\n",
        "\n",
        "Training the model with a Learning Rate of 0.05"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FcnqOwQv5cKK"
      },
      "outputs": [],
      "source": [
        "# define optimizer and loss function\n",
        "\n",
        "model2 = TorchConnector(qnn2, initial_weights)\n",
        "\n",
        "optimizer = LBFGS(model2.parameters(),lr=0.05)\n",
        "f_loss = CrossEntropyLoss()\n",
        "\n",
        "X= [normlaizeData(dataInput[i].flatten()) for i in range(50)]\n",
        "y01= [data_target_o[i] for i in range(50)]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "K0iuvuUi5cKM",
        "outputId": "47817de0-799c-4d9a-b9ed-2bf90f3d4060"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy: 0.46\n"
          ]
        }
      ],
      "source": [
        "from torch import Tensor\n",
        "# traning model accuracy\n",
        "y_predict = []\n",
        "for x in X:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy:', sum(y_predict == np.array(y01))/len(np.array(y01)))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "This accuracy without applying any optimizer and just running on the first instance, as expected that the model is correct half of the time, even if you randomly classify the images you'll get the accuracy near $50\\%$ , so that's what we see here."
      ],
      "metadata": {
        "id": "Fm1Jj2XKENP_"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wC9qiJET5cKO"
      },
      "outputs": [],
      "source": [
        "# define optimizer and loss function\n",
        "\n",
        "model2 = TorchConnector(qnn2, initial_weights)\n",
        "\n",
        "optimizer = LBFGS(model2.parameters(),lr=0.05)\n",
        "f_loss = CrossEntropyLoss()\n",
        "\n",
        "X= [normlaizeData(dataInput[i].flatten()) for i in range(50)]\n",
        "y01= [data_target_o[i] for i in range(50)]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LEQQKM__5cKQ",
        "outputId": "49afb082-5339-49e6-a86a-11a16d7c1282"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "33.18708801269531\n",
            "33.108192443847656\n",
            "32.4072380065918\n",
            "32.03237533569336\n",
            "31.895145416259766\n",
            "31.78937530517578\n",
            "31.689733505249023\n",
            "31.59193229675293\n",
            "31.493995666503906\n",
            "31.396451950073242\n",
            "31.304031372070312\n",
            "31.22391700744629\n",
            "31.15971565246582\n",
            "31.109495162963867\n",
            "31.069421768188477\n",
            "31.036239624023438\n",
            "31.007585525512695\n",
            "30.98176383972168\n",
            "30.957427978515625\n",
            "30.93332862854004\n",
            "30.908166885375977\n",
            "30.88027572631836\n",
            "30.84713363647461\n",
            "30.804162979125977\n",
            "30.741378784179688\n",
            "30.637073516845703\n",
            "30.515186309814453\n",
            "30.469327926635742\n",
            "30.444242477416992\n",
            "30.421152114868164\n",
            "30.39798355102539\n",
            "30.37508773803711\n",
            "30.352642059326172\n",
            "30.330595016479492\n",
            "30.308734893798828\n",
            "30.286935806274414\n",
            "30.26512908935547\n",
            "30.24332046508789\n",
            "30.221588134765625\n",
            "30.200000762939453\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "tensor(30.9082, grad_fn=<AddBackward0>)"
            ]
          },
          "execution_count": 10,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "from torch import Tensor\n",
        "# start training\n",
        "\n",
        "model2.train()    # set model to training mode\n",
        "\n",
        "# define objective function\n",
        "def closure():\n",
        "    optimizer.zero_grad()                                  # initialize gradient\n",
        "    loss = 0.0                                             # initialize loss    \n",
        "    for x, y_target in zip(X, y01):                        # evaluate batch loss\n",
        "        output = model2(Tensor(x)).reshape(1, 2)           # forward pass\n",
        "        loss += f_loss(output, Tensor([y_target]).long())\n",
        "    loss.backward()                                        # backward pass\n",
        "    print(loss.item())                                     # print loss\n",
        "    return loss\n",
        "\n",
        "# run optimizer\n",
        "optimizer.step(closure)\n",
        "optimizer.step(closure)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VWezyfvi5cKR",
        "outputId": "4aab0de1-7161-47eb-e6f4-69ecfef059e1"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy: 0.92\n"
          ]
        }
      ],
      "source": [
        "# traning model accuracy\n",
        "y_predict = []\n",
        "for x in X:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy:', sum(y_predict == np.array(y01))/len(np.array(y01)))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here you can see that the accuracy goes to $92\\%$ , which says that our model is good at classifying."
      ],
      "metadata": {
        "id": "JVM_b1C7EBZm"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x9ZsRoQh5cKV"
      },
      "source": [
        "# Learning rate 0.06\n",
        "\n",
        "Increasing the Learning Rate with $+0.01$ to $0.06$, we'll see how it affects the accuracy."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "pSBdJB1h5cKX",
        "outputId": "52bd2aa4-df87-4f4e-f376-ae513d0e5fe3"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy: 0.46\n"
          ]
        }
      ],
      "source": [
        "# define optimizer and loss function\n",
        "from torch import Tensor\n",
        "model2 = TorchConnector(qnn2, initial_weights)\n",
        "\n",
        "optimizer = LBFGS(model2.parameters(),lr=0.06)\n",
        "f_loss = CrossEntropyLoss()\n",
        "\n",
        "X= [normlaizeData(dataInput[i].flatten()) for i in range(50)]\n",
        "y01= [data_target_o[i] for i in range(50)]\n",
        "\n",
        "y_predict = []\n",
        "for x in X:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy:', sum(y_predict == np.array(y01))/len(np.array(y01)))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8CbTEQys5cKZ",
        "outputId": "af452188-8426-4ba7-e912-35cb101c0aaf"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "33.18708801269531\n",
            "33.092506408691406\n",
            "32.27847671508789\n",
            "31.932584762573242\n",
            "31.801755905151367\n",
            "31.696422576904297\n",
            "31.596323013305664\n",
            "31.49781036376953\n",
            "31.399715423583984\n",
            "31.304540634155273\n",
            "31.219013214111328\n",
            "31.14921760559082\n",
            "31.095157623291016\n",
            "31.052919387817383\n",
            "31.01851463317871\n",
            "30.989017486572266\n",
            "30.96230697631836\n",
            "30.936681747436523\n",
            "30.910457611083984\n",
            "30.88165283203125\n",
            "30.847291946411133\n",
            "30.801864624023438\n",
            "30.7325382232666\n",
            "30.610553741455078\n",
            "30.497982025146484\n",
            "30.462387084960938\n",
            "30.439420700073242\n",
            "30.41527557373047\n",
            "30.38955307006836\n",
            "30.363304138183594\n",
            "30.336950302124023\n",
            "30.31045150756836\n",
            "30.28369903564453\n",
            "30.256694793701172\n",
            "30.22960662841797\n",
            "30.202392578125\n",
            "30.172588348388672\n",
            "30.144634246826172\n",
            "30.120256423950195\n",
            "30.096738815307617\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "tensor(30.8473, grad_fn=<AddBackward0>)"
            ]
          },
          "execution_count": 13,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "from torch import Tensor\n",
        "# start training\n",
        "\n",
        "model2.train()    # set model to training mode\n",
        "\n",
        "# define objective function\n",
        "def closure():\n",
        "    optimizer.zero_grad()                                  # initialize gradient\n",
        "    loss = 0.0                                             # initialize loss    \n",
        "    for x, y_target in zip(X, y01):                        # evaluate batch loss\n",
        "        output = model2(Tensor(x)).reshape(1, 2)           # forward pass\n",
        "        loss += f_loss(output, Tensor([y_target]).long())\n",
        "    loss.backward()                                        # backward pass\n",
        "    print(loss.item())                                     # print loss\n",
        "    return loss\n",
        "\n",
        "# run optimizer\n",
        "optimizer.step(closure)\n",
        "optimizer.step(closure)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "I33pz30p5cKb",
        "outputId": "4c015516-4b15-4a64-ae4a-c0312e6c47ee"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy: 0.92\n"
          ]
        }
      ],
      "source": [
        "# traning model accuracy\n",
        "y_predict = []\n",
        "for x in X:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy:', sum(y_predict == np.array(y01))/len(np.array(y01)))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We see the accuracy drops to $92\\%$ from $94\\%$."
      ],
      "metadata": {
        "id": "qQUb7q5BEyqZ"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1co0XUZm5cKd"
      },
      "source": [
        "# Learning Rate 0.07\n",
        "\n",
        "Increasing the accuracy even further by $0.01$"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HkMzQ3VM5cKe",
        "outputId": "a96725e8-4003-4381-fd06-93abbbd5af38"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy: 0.46\n"
          ]
        }
      ],
      "source": [
        "# define optimizer and loss function\n",
        "from torch import Tensor\n",
        "model2 = TorchConnector(qnn2, initial_weights)\n",
        "\n",
        "optimizer = LBFGS(model2.parameters(),lr=0.07)\n",
        "f_loss = CrossEntropyLoss()\n",
        "\n",
        "X= [normlaizeData(dataInput[i].flatten()) for i in range(50)]\n",
        "y01= [data_target_o[i] for i in range(50)]\n",
        "\n",
        "y_predict = []\n",
        "for x in X:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy:', sum(y_predict == np.array(y01))/len(np.array(y01)))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "43MQltGY5cKg",
        "outputId": "2e02fbdc-8f77-4281-8a9d-f15f1ebdc4bc"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "33.18708801269531\n",
            "33.07685089111328\n",
            "32.15932846069336\n",
            "31.843116760253906\n",
            "31.71990394592285\n",
            "31.616920471191406\n",
            "31.518083572387695\n",
            "31.42047691345215\n",
            "31.324111938476562\n",
            "31.233501434326172\n",
            "31.156253814697266\n",
            "31.095962524414062\n",
            "31.04983901977539\n",
            "31.013206481933594\n",
            "30.98231315612793\n",
            "30.954456329345703\n",
            "30.927480697631836\n",
            "30.899188995361328\n",
            "30.866764068603516\n",
            "30.82535743713379\n",
            "30.764251708984375\n",
            "30.654775619506836\n",
            "30.505294799804688\n",
            "30.467437744140625\n",
            "30.445926666259766\n",
            "30.420310974121094\n",
            "30.390424728393555\n",
            "30.358802795410156\n",
            "30.326568603515625\n",
            "30.2938232421875\n",
            "30.260597229003906\n",
            "30.227100372314453\n",
            "30.19232749938965\n",
            "30.15595245361328\n",
            "30.12395477294922\n",
            "30.094696044921875\n",
            "30.066675186157227\n",
            "30.039791107177734\n",
            "30.01435661315918\n",
            "29.9906005859375\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "tensor(30.7643, grad_fn=<AddBackward0>)"
            ]
          },
          "execution_count": 16,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "from torch import Tensor\n",
        "# start training\n",
        "\n",
        "model2.train()    # set model to training mode\n",
        "\n",
        "# define objective function\n",
        "def closure():\n",
        "    optimizer.zero_grad()                                  # initialize gradient\n",
        "    loss = 0.0                                             # initialize loss    \n",
        "    for x, y_target in zip(X, y01):                        # evaluate batch loss\n",
        "        output = model2(Tensor(x)).reshape(1, 2)           # forward pass\n",
        "        loss += f_loss(output, Tensor([y_target]).long())\n",
        "    loss.backward()                                        # backward pass\n",
        "    print(loss.item())                                     # print loss\n",
        "    return loss\n",
        "\n",
        "# run optimizer\n",
        "optimizer.step(closure)\n",
        "optimizer.step(closure)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ge1Gn9LZ5cKi",
        "outputId": "130750ae-3ca5-48b4-b17a-151f361b31bc"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy: 0.92\n"
          ]
        }
      ],
      "source": [
        "# traning model accuracy\n",
        "y_predict = []\n",
        "for x in X:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy:', sum(y_predict == np.array(y01))/len(np.array(y01)))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The acciracy remained constant at $92\\%$"
      ],
      "metadata": {
        "id": "gqx8ign8FBr8"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5S5UyxQK5cKq"
      },
      "source": [
        "# Test Model\n",
        "\n",
        "Testing the model with unseen data."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fVScoBD75cKs",
        "outputId": "2f0f203e-74e0-4fa1-d36c-c54de5b65489"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy25data: 0.84\n",
            "Accuracy50data: 0.78\n"
          ]
        }
      ],
      "source": [
        "target_o = [1 for i in range(25)]+[0 for i in range(25)]\n",
        "\n",
        "pathY=r'../dataset/Original/galaxy1/'\n",
        "pathN=r'../dataset/Original/no-galaxy1/'\n",
        "nameN=''\n",
        "nameY=''\n",
        "inputY=[imageResize(callImage(i+1,pathY,nameY),16) for i in range(25)]\n",
        "inputN=[imageResize(callImage(i+1,pathN,nameN),16) for i in range(25)]\n",
        "input_combine = inputY+inputN\n",
        "\n",
        "np.random.seed(0)\n",
        "idx=np.array([int(i) for i in range(50)]).flatten()\n",
        "\n",
        "np.random.shuffle(idx)\n",
        "\n",
        "dataInput = list(input_combine[i] for i in idx )\n",
        "dataTarget = list( imageBinarize(input_combine[i]) for i in idx )\n",
        "\n",
        "data_target_o=list( target_o[i] for i in idx )\n",
        "\n",
        "Xtest= [normlaizeData(dataInput[i].flatten()) for i in range(25)]\n",
        "y01test= [data_target_o[i] for i in range(25)]\n",
        "\n",
        "Xtest1= [normlaizeData(dataInput[i].flatten()) for i in range(50)]\n",
        "y01test1= [data_target_o[i] for i in range(50)]\n",
        "\n",
        "y_predict = []\n",
        "for x in Xtest:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy25data:', sum(y_predict == np.array(y01test))/len(np.array(y01test)))\n",
        "\n",
        "y_predict1 = []\n",
        "for x in Xtest1:\n",
        "    output = model2(Tensor(x))\n",
        "    y_predict1 += [np.argmax(output.detach().numpy())]\n",
        "\n",
        "print('Accuracy50data:', sum(y_predict1 == np.array(y01test1))/len(np.array(y01test1)))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The accuracy on different unseen data ranges from $80-84\\%$.\n",
        "\n",
        "\n",
        "We can say the model performs good on the unseen data too.\n",
        " \n",
        "\n",
        "We need new techniques to train the Hyperparameters to better fine tune them and reduce the computational time."
      ],
      "metadata": {
        "id": "O6HXVKzwFU9t"
      }
    }
  ],
  "metadata": {
    "interpreter": {
      "hash": "3bc52ef4732c2c89ac1c67f1a564bc39cd50f6fed15c43a2bb566bf993475a13"
    },
    "kernelspec": {
      "display_name": "Python 3.8.9 ('qc': venv)",
      "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": "3.8.9"
    },
    "colab": {
      "provenance": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}