patchcore_run_notebook.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "import warnings  \n",
        "from data_loading import MVTecDataset, _CLASSNAMES, _CLASSNAMES_1, _CLASSNAMES_2, _CLASSNAMES_3, _CLASSNAMES_4\n",
        "from patchcore_model import PatchCore\n",
        "from utils import print_results\n",
        "\n",
        "from typing import List\n",
        "import torch\n",
        "import random\n",
        "import numpy as np\n",
        "torch.manual_seed(0)\n",
        "random.seed(0)\n",
        "np.random.seed(0)\n",
        "\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "\n",
        "ALL_CLASSES = _CLASSNAMES\n",
        "BATCH_CLASSES_1 = _CLASSNAMES_1\n",
        "BATCH_CLASSES_2 = _CLASSNAMES_2 \n",
        "BATCH_CLASSES_3 = _CLASSNAMES_3\n",
        "BATCH_CLASSES_4 = _CLASSNAMES_4\n",
        "test = [\"bottle\"]"
      ],
      "metadata": {
        "id": "v8vahNC3AdgA"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def create_model(model_name: str):\n",
        "    if model_name == \"PatchCore\":\n",
        "        return PatchCore(f_coreset=.25, backbone_name=\"wideresnet50\")\n",
        "    else:\n",
        "        raise ValueError(f\"The model {model_name} do not exists or is not avaiable here\")\n",
        "\n",
        "def train_model(model, train_dataset):\n",
        "    model.fit(train_dataset)\n",
        "\n",
        "def evaluate_model(model, test_dataset, cls):\n",
        "    return model.evaluate(test_dataset, cls = cls)\n",
        "\n",
        "# Main method that will allow running the program\n",
        "def run_model(model_name: str, classes: List):\n",
        "\n",
        "    results = {}\n",
        "\n",
        "\n",
        "    nb_datasets = len(classes)\n",
        "\n",
        "    # This function iterates over the different datasets to train the model on it, test it with the test images and evaluate the performance\n",
        "    for cls in classes:\n",
        "\n",
        "        # Creation of the model\n",
        "        model = create_model(model_name)\n",
        "        print(f\"\\nThe model is runing on the {cls} dataset ...\")\n",
        "\n",
        "        # Load of the data related to the current dataset\n",
        "        train_ds, test_ds = MVTecDataset(cls).get_dataloaders()\n",
        "\n",
        "        # Train the model using the training data\n",
        "        train_model(model, train_ds)\n",
        "\n",
        "        # Use the test data to apply the trained model on it, try to classify it and evaluate the performance\n",
        "        image_rocauc = evaluate_model(model, test_ds, cls)\n",
        "        \n",
        "        print(\"Here are the test results on the dataset\", cls)\n",
        "        print(\"ROC_AUC metric :\", image_rocauc)\n",
        "\n",
        "        results[cls] = [float(image_rocauc)]\n",
        "        image_results = [v[0] for _, v in results.items()]\n",
        "\n",
        "        # Compute the average score based on the scores obtained on the dataset tested\n",
        "        average_image_roc_auc = sum(image_results)/len(image_results)\n",
        "\n",
        "\n",
        "        total_results = {\n",
        "            \"per_class_results\": results,\n",
        "            \"average image rocauc\": average_image_roc_auc,\n",
        "            \"model parameters\": model.get_parameters(),\n",
        "        }\n",
        "\n",
        "    return \"After \" + str(nb_datasets) + \" iterations performed on different datasets, here is the final average results obtained :\" + str(total_results)\n",
        "\n",
        "\n",
        "run_model(\"PatchCore\", BATCH_CLASSES_1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "szRykBJ7Aeal",
        "outputId": "32029f11-cae9-4891-f368-f1d57951099c"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "The model is runing on the bottle dataset ...\n",
            "Beginning of the fit method...\n",
            "Beginning of the coreset subsampling reduction...\n",
            "   Fitting random projections. Start dim = torch.Size([163856, 1536]).\n",
            "   DONE.                 Transformed dim = torch.Size([163856, 296]).\n",
            "End of the coreset subsampling reduction\n",
            "Sklearn roc_auc metric computation loading ...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAdx0lEQVR4nO3de5xVdb3/8ddbQQflUgfGQi6BOZJ4CXXykh6VSCUvoD9JMc3omJRmmmiPHx3L8nI8eKgsz0ETzVDD8cI5CRWmZpBm3gBRBFIIQQZQCcjLQRT0c/5Ya2g7zGUPM2vvZtb7+Xjsx6zLd6/1+c7ts7/ru9b3q4jAzMzya4dyB2BmZuXlRGBmlnNOBGZmOedEYGaWc04EZmY516ncAbRUr169YsCAAeUOw8ysXZk7d+5fI6KyoX3tLhEMGDCAOXPmlDsMM7N2RdKKxvb50pCZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOZZYIJN0q6TVJzzeyX5Kul7RU0nOSDswqFjMza1yWLYIpwPAm9n8OqEpfY4EbM4zFzMwakdlzBBHxiKQBTRQZCdweyTjYT0j6kKTeEbEmi3jufPJlps9flcWhzcxKYvDu3fneSfu0+XHL2UfQB1hZsF6bbtuGpLGS5kias3bt2u062fT5q1i05o3teq+ZWUfWLp4sjojJwGSA6urq7Z5JZ3Dv7tz91cPaLC4zs46gnC2CVUC/gvW+6TYzMyuhciaCGcDZ6d1DhwKvZ9U/YGZmjcvs0pCkGuBooJekWuB7QGeAiPgpMBM4HlgKbAS+nFUsZmbWuCzvGjqjmf0BfD2r85uZWXH8ZLGZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOORGYmeWcE4GZWc5lmggkDZf0gqSlksY3sL+/pFmSnpH0nKTjs4zHzMy2lVkikLQjMAn4HDAYOEPS4HrFvgPcExEHAKOBG7KKx8zMGpZli+BgYGlELIuId4G7gJH1ygTQPV3uAazOMB4zM2tAlomgD7CyYL023Vbo+8BZkmqBmcA3GjqQpLGS5kias3bt2ixiNTPLrXJ3Fp8BTImIvsDxwB2StokpIiZHRHVEVFdWVpY8SDOzjizLRLAK6Few3jfdVugc4B6AiHgcqAB6ZRiTmZnVk2UieBqokjRQ0k4kncEz6pV5GRgGIGlvkkTgaz9mZiWUWSKIiC3ABcADwGKSu4MWSrpS0oi02CXAuZKeBWqAMRERWcVkZmbb6pTlwSNiJkkncOG2ywuWFwGHZxmDmZk1rdydxWZmVmZOBGZmOedEYGaWc04EZmY550RgZpZzTgRmZjnnRGBmlnNOBGZmOedEYGaWc04EZmY5V3QikLRLloGYmVl5NJsIJH1a0iLgz+n6JyV5Skkzsw6imBbBdcBxwDqAiHgWODLLoMzMrHSKujQUESvrbXovg1jMzKwMihmGeqWkTwMhqTNwEcn8AmZm1gEU0yL4GvB1konnVwFDgPOzDMrMzEqnmBbBoIg4s3CDpMOBx7IJyczMSqmYFsF/FrnNzMzaoUZbBJIOAz4NVEoaV7CrO7Bj1oGZmVlpNHVpaCega1qmW8H2N4BRWQZlZmal02giiIg/AH+QNCUiVpQwJjMzK6FiOos3SpoI7ANU1G2MiM9kFpWZmZVMMZ3FU0mGlxgIXAEsB57OMCYzMyuhYhJBz4j4GbA5Iv4QEf8CuDVgZtZBFHNpaHP6dY2kE4DVwD9lF5KZmZVSMYngakk9gEtInh/oDnwz06jMzKxkmk0EEfHrdPF1YChsfbLYzMw6gKYeKNsROI1kjKHfRsTzkk4E/hXoAhxQmhDNzCxLTbUIfgb0A54Crpe0GqgGxkfEfaUIzszMstdUIqgG9o+I9yVVAK8AH4+IdaUJzczMSqGp20ffjYj3ASJiE7CspUlA0nBJL0haKml8I2VOk7RI0kJJd7bk+GZm1npNtQg+Iem5dFnAx9N1ARER+zd14LSPYRJwDFALPC1pRkQsKihTBXwbODwiNkjarRV1MTOz7dBUIti7lcc+GFgaEcsAJN0FjAQWFZQ5F5gUERsAIuK1Vp7TzMxaqKlB51o70FwfoHCu41rgkHpl9gKQ9BjJ0Nbfj4jf1j+QpLHAWID+/fu3MiwzMytU1OT1GeoEVAFHA2cAN0v6UP1CETE5IqojorqysrLEIZqZdWxZJoJVJLef1umbbitUC8yIiM0R8RLwIkliMDOzEikqEUjqImlQC4/9NFAlaaCknYDRwIx6Ze4jaQ0gqRfJpaJlLTyPmZm1QrOJQNJJwHzgt+n6EEn1/6FvIyK2ABcADwCLgXsiYqGkKyWNSIs9AKyTtAiYBXzLzymYmZVWMYPOfZ/kDqDZABExX9LAYg4eETOBmfW2XV6wHMC49GVmZmVQzKWhzRHxer1tkUUwZmZWesW0CBZK+gKwY/oA2IXAn7INy8zMSqWYFsE3SOYrfge4k2Q4as9HYGbWQRTTIvhERFwGXJZ1MGZmVnrFtAh+KGmxpKsk7Zt5RGZmVlLNJoKIGEoyM9la4CZJCyR9J/PIzMysJIp6oCwiXomI64GvkTxTcHkzbzEzs3aimAfK9pb0fUkLSCav/xPJcBFmZtYBFNNZfCtwN3BcRKzOOB4zMyuxZhNBRBxWikDMzKw8Gk0Eku6JiNPSS0KFTxIXNUOZmZm1D021CC5Kv55YikDMzKw8Gu0sjog16eL5EbGi8AWcX5rwzMwsa8XcPnpMA9s+19aBmJlZeTTVR3AeySf/PSQ9V7CrG/BY1oGZmVlpNNVHcCdwP/DvwPiC7W9GxPpMozIzs5JpKhFERCyX9PX6OyT9k5OBmVnH0FyL4ERgLsntoyrYF8AeGcZlZmYl0mgiiIgT069FTUtpZmbtUzFjDR0uadd0+SxJP5LUP/vQzMysFIq5ffRGYKOkTwKXAH8B7sg0KjMzK5liEsGWiAhgJPBfETGJ5BZSMzPrAIoZffRNSd8Gvgj8s6QdgM7ZhmVmZqVSTIvgdJKJ6/8lIl4hmYtgYqZRmZlZyRQzVeUrwFSgh6QTgU0RcXvmkZmZWUkUc9fQacBTwOeB04AnJY3KOjAzMyuNYvoILgM+FRGvAUiqBH4HTMsyMDMzK41i+gh2qEsCqXVFvs/MzNqBYloEv5X0AFCTrp8OzMwuJDMzK6Vi5iz+lqT/BxyRbpocEb/MNiwzMyuVpuYjqAJ+AHwcWABcGhGrShWYmZmVRlPX+m8Ffg2cSjIC6X+29OCShkt6QdJSSeObKHeqpJBU3dJzmJlZ6zR1aahbRNycLr8gaV5LDixpR2ASyVSXtcDTkmZExKJ65boBFwFPtuT4ZmbWNppKBBWSDuDv8xB0KVyPiOYSw8HA0ohYBiDpLpLxihbVK3cVcC3wrRbGbmZmbaCpRLAG+FHB+isF6wF8pplj9wFWFqzXAocUFpB0INAvIn4jqdFEIGksMBagf3+PgG1m1paamphmaJYnTgev+xEwprmyETEZmAxQXV0dWcZlZpY3WT4YtgroV7DeN91WpxuwLzBb0nLgUGCGO4zNzEory0TwNFAlaaCknYDRwIy6nRHxekT0iogBETEAeAIYERFzMozJzMzqySwRRMQW4ALgAWAxcE9ELJR0paQRWZ3XzMxaptkniyUJOBPYIyKuTOcr/mhEPNXceyNiJvWGo4iIyxspe3RREZuZWZsqpkVwA3AYcEa6/ibJ8wFmZtYBFDPo3CERcaCkZwAiYkN6zd/MzDqAYloEm9OnhAO2zkfwfqZRmZlZyRSTCK4HfgnsJunfgD8C12QalZmZlUwxw1BPlTQXGEYyvMTJEbE488jMzKwkirlrqD+wEfhV4baIeDnLwMzMrDSK6Sz+DUn/gIAKYCDwArBPhnGZmVmJFHNpaL/C9XSguPMzi8jMzEqqxU8Wp8NPH9JsQTMzaxeK6SMYV7C6A3AgsDqziMzMrKSK6SPoVrC8haTP4L+zCcfMzEqtyUSQPkjWLSIuLVE8ZmZWYo32EUjqFBHvAYeXMB4zMyuxploET5H0B8yXNAO4F/jfup0R8T8Zx2ZmZiVQTB9BBbCOZI7iuucJAnAiMDPrAJpKBLuldww9z98TQB3PG2xm1kE0lQh2BLrywQRQx4nAzKyDaCoRrImIK0sWiZmZlUVTTxY31BIwM7MOpqlEMKxkUZiZWdk0mggiYn0pAzEzs/Jo8aBzZmbWsTgRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOZZoIJA2X9IKkpZLGN7B/nKRFkp6T9LCkj2UZj5mZbSuzRJDOdzwJ+BwwGDhD0uB6xZ4BqiNif2Aa8B9ZxWNmZg3LskVwMLA0IpZFxLvAXcDIwgIRMSsiNqarTwB9M4zHzMwakGUi6AOsLFivTbc15hzg/oZ2SBoraY6kOWvXrm3DEM3M7B+is1jSWUA1MLGh/RExOSKqI6K6srKytMGZmXVwxUxev71WAf0K1vum2z5A0meBy4CjIuKdDOMxM7MGZNkieBqokjRQ0k7AaGBGYQFJBwA3ASMi4rUMYzEzs0ZklggiYgtwAfAAsBi4JyIWSrpS0oi02ESgK3CvpPmSZjRyODMzy0iWl4aIiJnAzHrbLi9Y/myW5zczs+b9Q3QWm5lZ+TgRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOORGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkRmJnlnBOBmVnOdSp3AGa2/TZv3kxtbS2bNm0qdyj2D6KiooK+ffvSuXPnot/jRGDWjtXW1tKtWzcGDBiApHKHY2UWEaxbt47a2loGDhxY9Pt8acisHdu0aRM9e/Z0EjAAJNGzZ88WtxCdCMzaOScBK7Q9vw9OBGZmOedEYGbbbfny5ey7774tes+Pf/xjNm7cuHX9mmuu+cD+rl27tuh4EcFnPvMZ3njjja3b7rvvPiTx5z//eeu22bNnc+KJJ37gvWPGjGHatGlA0vE+fvx4qqqqOPDAAznssMO4//77WxRLfevWrWPo0KF07dqVCy64oNFy69ev55hjjqGqqopjjjmGDRs2bK3bhRdeyJ577sn+++/PvHnzAFi7di3Dhw9vVWyFnAjMrKSaSwQtNXPmTD75yU/SvXv3rdtqamo44ogjqKmpKfo43/3ud1mzZg3PP/888+bN47777uPNN99sVWwVFRVcddVV/OAHP2iy3IQJExg2bBhLlixh2LBhTJgwAYD777+fJUuWsGTJEiZPnsx5550HQGVlJb179+axxx5rVXx1fNeQWQdxxa8Wsmj1G80XbIHBu3fneyft02SZLVu2cOaZZzJv3jz22Wcfbr/9dnbZZRcefvhhLr30UrZs2cKnPvUpbrzxRm666SZWr17N0KFD6dWrF4cccghvv/02Q4YMYZ999mHq1KkfOPbEiRO55557eOeddzjllFO44oortjn/1KlTGTt27Nb1t956iz/+8Y/MmjWLk046qcH31Ldx40ZuvvlmXnrpJXbeeWcAPvKRj3DaaacV821q1K677soRRxzB0qVLmyw3ffp0Zs+eDcCXvvQljj76aK699lqmT5/O2WefjSQOPfRQ/va3v7FmzRp69+7NySefzNSpUzn88MNbFSO4RWBmrfTCCy9w/vnns3jxYrp3784NN9zApk2bGDNmDHfffTcLFixgy5Yt3HjjjVx44YXsvvvuzJo1i1mzZjFhwgS6dOnC/Pnzt0kCDz74IEuWLOGpp55i/vz5zJ07l0ceeWSb8z/22GMcdNBBW9enT5/O8OHD2WuvvejZsydz585ttg5Lly6lf//+H2hVNObiiy9myJAh27zqPsVvj1dffZXevXsD8NGPfpRXX30VgFWrVtGvX7+t5fr27cuqVasAqK6u5tFHH93ucxZyi8Csg2juk3tW+vXrt/VT6VlnncX111/PMcccw8CBA9lrr72A5FPupEmT+OY3v1n0cR988EEefPBBDjjgACD5pL9kyRKOPPLID5Rbv3493bp127peU1PDRRddBMDo0aOpqanhoIMOavRumpbeZXPddde1qHxLSSoqpt12243Vq1e3yTkzTQSShgM/AXYEbomICfX27wzcDhwErANOj4jlWcZkZm2r/j+ttrqdNSL49re/zVe/+tUmy3Xq1In333+fHXbYgfXr1/P73/+eBQsWIIn33nsPSUycOJGePXtu7YSts379enr16sWee+7Jyy+/zBtvvNFsq+Diiy9m1qxZ22wfPXo048ePb3lFSS5D1V3yWbNmDbvtthsAffr0YeXKlVvL1dbW0qdPHyB5hqRLly7bdb76Mrs0JGlHYBLwOWAwcIakwfWKnQNsiIg9geuAa7OKx8yy8fLLL/P4448DcOedd3LEEUcwaNAgli9fvvXa+B133MFRRx0FQLdu3T7QCdu5c2c2b968zXGPO+44br31Vt566y0guUzy2muvbVNu0KBBLFu2DIBp06bxxS9+kRUrVrB8+XJWrlzJwIEDefTRR6mqqmL16tUsXrwYgBUrVvDss88yZMgQdtllF8455xwuuugi3n33XSC5M+fee+/d5nzXXXcd8+fP3+a1vUkAYMSIEdx2220A3HbbbYwcOXLr9ttvv52I4IknnqBHjx5bLyG9+OKLLb5jqzFZ9hEcDCyNiGUR8S5wFzCyXpmRwG3p8jRgmPx0jFm7MmjQICZNmsTee+/Nhg0bOO+886ioqODnP/85n//859lvv/3YYYcd+NrXvgbA2LFjGT58OEOHDt26vv/++3PmmWd+4LjHHnssX/jCFzjssMPYb7/9GDVqVIN38ZxwwglbO1pramo45ZRTPrD/1FNPpaamhp133plf/OIXfPnLX2bIkCGMGjWKW265hR49egBw9dVXU1lZyeDBg9l333058cQTi+ozaM6AAQMYN24cU6ZMoW/fvixatAiAr3zlK8yZMweA8ePH89BDD1FVVcXvfve7rUnl+OOPZ4899mDPPffk3HPP5YYbbth63FmzZnHCCSe0Oj4ARUSbHGibA0ujgOER8ZV0/YvAIRFxQUGZ59Myten6X9Iyf613rLHAWID+/fsftGLFihbHc8WvFgLlu45qloXFixez9957lzuMslqzZg1nn302Dz30ULlDKakjjzyS6dOn8+EPf3ibfQ39XkiaGxHVDR2rXXQWR8RkYDJAdXX1dmUuJwCzjql3796ce+65RV3f7yjWrl3LuHHjGkwC2yPLRLAK6Few3jfd1lCZWkmdgB4kncZmZkVr7f3+7U1lZSUnn3xymx0vyz6Cp4EqSQMl7QSMBmbUKzMD+FK6PAr4fWR1rcqsg/KfjBXant+HzBJBRGwBLgAeABYD90TEQklXShqRFvsZ0FPSUmAcsP3d7mY5VFFRwbp165wMDPj7fAQVFRUtel9mncVZqa6ujrqedrO88wxlVl9jM5S1+85iM2tY586dWzQTlVlDPNaQmVnOORGYmeWcE4GZWc61u85iSWuBlj9anOgF/LXZUh2L65wPrnM+tKbOH4uIyoZ2tLtE0BqS5jTWa95Ruc754DrnQ1Z19qUhM7OccyIwM8u5vCWCyeUOoAxc53xwnfMhkzrnqo/AzMy2lbcWgZmZ1eNEYGaWcx0yEUgaLukFSUslbTOiqaSdJd2d7n9S0oDSR9m2iqjzOEmLJD0n6WFJHytHnG2puToXlDtVUkhq97caFlNnSaelP+uFku4sdYxtrYjf7f6SZkl6Jv39Pr4ccbYVSbdKei2dwbGh/ZJ0ffr9eE7Sga0+aUR0qBewI/AXYA9gJ+BZYHC9MucDP02XRwN3lzvuEtR5KLBLunxeHuqclusGPAI8AVSXO+4S/JyrgGeAD6fru5U77hLUeTJwXro8GFhe7rhbWecjgQOB5xvZfzxwPyDgUODJ1p6zI7YIDgaWRsSyiHgXuAsYWa/MSOC2dHkaMEySShhjW2u2zhExKyI2pqtPkMwY154V83MGuAq4FugI4zQXU+dzgUkRsQEgIl4rcYxtrZg6B1A3R2UPYHUJ42tzEfEIsL6JIiOB2yPxBPAhSb1bc86OmAj6ACsL1mvTbQ2WiWQCndeBniWJLhvF1LnQOSSfKNqzZuucNpn7RcRvShlYhor5Oe8F7CXpMUlPSBpesuiyUUydvw+cJakWmAl8ozShlU1L/96b5fkIckbSWUA1cFS5Y8mSpB2AHwFjyhxKqXUiuTx0NEmr7xFJ+0XE38oaVbbOAKZExA8lHQbcIWnfiHi/3IG1Fx2xRbAK6Few3jfd1mAZSZ1ImpPrShJdNoqpM5I+C1wGjIiId0oUW1aaq3M3YF9gtqTlJNdSZ7TzDuNifs61wIyI2BwRLwEvkiSG9qqYOp8D3AMQEY8DFSSDs3VURf29t0RHTARPA1WSBkraiaQzeEa9MjOAL6XLo4DfR9oL0041W2dJBwA3kSSB9n7dGJqpc0S8HhG9ImJARAwg6RcZERHteZ7TYn637yNpDSCpF8mlomWlDLKNFVPnl4FhAJL2JkkEa0saZWnNAM5O7x46FHg9Ita05oAd7tJQRGyRdAHwAMkdB7dGxEJJVwJzImIG8DOS5uNSkk6Z0eWLuPWKrPNEoCtwb9ov/nJEjChb0K1UZJ07lCLr/ABwrKRFwHvAtyKi3bZ2i6zzJcDNki4m6Tge054/2EmqIUnmvdJ+j+8BnQEi4qck/SDHA0uBjcCXW33Odvz9MjOzNtARLw2ZmVkLOBGYmeWcE4GZWc45EZiZ5ZwTgZlZzjkR2D8kSe9Jml/wGtBE2bfa4HxTJL2Unmte+oRqS49xi6TB6fK/1tv3p9bGmB6n7vvyvKRfSfpQM+WHtPfROC17vn3U/iFJeisiurZ12SaOMQX4dURMk3Qs8IOI2L8Vx2t1TM0dV9JtwIsR8W9NlB9DMurqBW0di3UcbhFYuyCpazqPwjxJCyRtM9KopN6SHin4xPzP6fZjJT2evvdeSc39g34E2DN977j0WM9L+ma6bVdJv5H0bLr99HT7bEnVkiYAXdI4pqb73kq/3iXphIKYp0gaJWlHSRMlPZ2OMf/VIr4tj5MONibp4LSOz0j6k6RB6ZO4VwKnp7GcnsZ+q6Sn0rINjdhqeVPusbf98quhF8lTsfPT1y9JnoLvnu7rRfJUZV2L9q306yXAZenyjiTjDfUi+ce+a7r9/wOXN3C+KcCodPnzwJPAQcACYFeSp7IXAgcApwI3F7y3R/p1NumcB3UxFZSpi/EU4LZ0eSeSUSS7AGOB76TbdwbmAAMbiPOtgvrdCwxP17sDndLlzwL/nS6PAf6r4P3XAGelyx8iGYto13L/vP0q76vDDTFhHcbbETGkbkVSZ+AaSUcC75N8Ev4I8ErBe54Gbk3L3hcR8yUdRTJZyWPp0Bo7kXySbshESd8hGafmHJLxa34ZEf+bxvA/wD8DvwV+KOlakstJj7agXvcDP5G0MzAceCQi3k4vR+0vaVRargfJYHEv1Xt/F0nz0/ovBh4qKH+bpCqSYRY6N3L+Y4ERki5N1yuA/umxLKecCKy9OBOoBA6KiM1KRhStKCwQEY+kieIEYIqkHwEbgIci4owizvGtiJhWtyJpWEOFIuJFJXMdHA9cLenhiLiymEpExCZJs4HjgNNJJlqBZLapb0TEA80c4u2IGCJpF5Lxd74OXE8yAc+siDgl7Vif3cj7BZwaES8UE6/lg/sIrL3oAbyWJoGhwDZzLiuZh/nViLgZuIVkur8ngMMl1V3z31XSXkWe81HgZEm7SNqV5LLOo5J2BzZGxC9IBvNraM7YzWnLpCF3kwwUVte6gOSf+nl175G0V3rOBkUy29yFwCX6+1DqdUMRjyko+ibJJbI6DwDfUNo8UjIqreWcE4G1F1OBakkLgLOBPzdQ5mjgWUnPkHza/klErCX5x1gj6TmSy0KfKOaEETGPpO/gKZI+g1si4hlgP+Cp9BLN94CrG3j7ZOC5us7ieh4kmRjod5FMvwhJ4loEzFMyaflNNNNiT2N5jmRilv8A/j2te+H7ZgGD6zqLSVoOndPYFqbrlnO+fdTMLOfcIjAzyzknAjOznHMiMDPLOScCM7OccyIwM8s5JwIzs5xzIjAzy7n/A8f0nPArBS3OAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Here are the test results on the dataset bottle\n",
            "ROC_AUC metric : 1.0\n",
            "\n",
            "The model is runing on the cable dataset ...\n",
            "The dataset cable is not already present in C:\\mvtec_anomaly_detection . Downloading cable ...\n",
            "Extracting the compressed folder cable .tar.xz....\n",
            "Download of cable in C:\\mvtec_anomaly_detection Complete.\n",
            "Beginning of the fit method...\n",
            "Beginning of the coreset subsampling reduction...\n",
            "   Fitting random projections. Start dim = torch.Size([175616, 1536]).\n",
            "   DONE.                 Transformed dim = torch.Size([175616, 298]).\n",
            "End of the coreset subsampling reduction\n",
            "Sklearn roc_auc metric computation loading ...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Here are the test results on the dataset cable\n",
            "ROC_AUC metric : 0.9672038980509745\n",
            "\n",
            "The model is runing on the capsule dataset ...\n",
            "The dataset capsule is not already present in C:\\mvtec_anomaly_detection . Downloading capsule ...\n",
            "Extracting the compressed folder capsule .tar.xz....\n",
            "Download of capsule in C:\\mvtec_anomaly_detection Complete.\n",
            "Beginning of the fit method...\n",
            "Beginning of the coreset subsampling reduction...\n",
            "   Fitting random projections. Start dim = torch.Size([171696, 1536]).\n",
            "   DONE.                 Transformed dim = torch.Size([171696, 297]).\n",
            "End of the coreset subsampling reduction\n",
            "Sklearn roc_auc metric computation loading ...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Here are the test results on the dataset capsule\n",
            "ROC_AUC metric : 0.973274830474671\n",
            "\n",
            "The model is runing on the carpet dataset ...\n",
            "The dataset carpet is not already present in C:\\mvtec_anomaly_detection . Downloading carpet ...\n",
            "Extracting the compressed folder carpet .tar.xz....\n",
            "Download of carpet in C:\\mvtec_anomaly_detection Complete.\n",
            "Beginning of the fit method...\n",
            "Beginning of the coreset subsampling reduction...\n",
            "   Fitting random projections. Start dim = torch.Size([219520, 1536]).\n",
            "   DONE.                 Transformed dim = torch.Size([219520, 303]).\n",
            "End of the coreset subsampling reduction\n",
            "Sklearn roc_auc metric computation loading ...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Here are the test results on the dataset carpet\n",
            "ROC_AUC metric : 0.9879614767255216\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "\"After 4 iterations performed on different datasets, here is the final average results obtained :{'per_class_results': {'bottle': [1.0], 'cable': [0.9672038980509745], 'capsule': [0.973274830474671], 'carpet': [0.9879614767255216]}, 'average image rocauc': 0.9821100513127918, 'model parameters': {'backbone_name': 'wideresnet50', 'out_indices': (2, 3), 'f_coreset': 0.25, 'n_reweight': 3}}\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "-- **SUMMARY OF THE OBTAINED RESULTS** -- \n",
        "\n",
        "Several results are displayed below and represent the roc_auc score computed for each dataset. The differents columns represent the *f_coreset* value used this the execution of the PatchCore algorithm. The first one corresponds to a value of **0.1**, which **10% of the memory bank** is kept during the **subsampling part**. The second one corresponds to a value of **0.25**, representing **25% of the memory bank**. \n",
        "\n",
        "```\n",
        "ROC_AUC score res| f_coreset :      0.1                0.25\n",
        "- bottle dataset              1.0                  1.0\n",
        "- cable dataset               0.9565217391304348   0.9636431784107946\n",
        "- capsule dataset             0.9792580773833267   0.9748703629836457\n",
        "- carpet dataset              0.9851524879614767   0.985553772070626\n",
        "- grid dataset                0.9598997493734336   0.9573934837092732\n",
        "- leather dataset             1.0                  1.0\n",
        "- metal_nut dataset           0.9828934506353861   0.9853372434017595\n",
        "- pill dataset                0.9399890889252592   0.92253136933988\n",
        "- screw dataset               0.9204755072760812   0.9020291043246567\n",
        "- tile dataset                0.9985569985569986   0.998917748917749\n",
        "- toothbrush dataset          0.9222222222222223   0.9222222222222223\n",
        "- transistor dataset          0.9954166666666666   0.9937499999999999\n",
        "- wood dataset                0.9885964912280701   0.9842105263157894\n",
        "- zipper dataset              0.9634978991596639   0.9663865546218487\n",
        "```\n",
        "\n",
        "Now let's retrieve these values to perform some statistic and visualization."
      ],
      "metadata": {
        "id": "VvQdQsYkXiCX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "-- **AVERAGE SCORE**--"
      ],
      "metadata": {
        "id": "arBtjrCCrDuI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# List of the roc_auc score obtained with f_coreset = 0.1\n",
        "results_fcor_0_1 = [1.0, 0.9565217391304348, 0.9792580773833267, 0.9851524879614767, 0.9598997493734336, 1.0, 0.9828934506353861,\n",
        "           0.9399890889252592, 0.9204755072760812, 0.9985569985569986, 0.9222222222222223, 0.9954166666666666, 0.9885964912280701, 0.963497899159663]\n",
        "\n",
        "# List of the roc_auc score obtained with f_coreset = 0.25\n",
        "results_fcor_0_25 = [1.0, 0.9636431784107946, 0.9748703629836457, 0.985553772070626, 0.9573934837092732, 1.0, 0.9853372434017595, \n",
        "           0.92253136933988, 0.9020291043246567, 0.998917748917749, 0.9222222222222223, 0.9937499999999999, 0.9842105263157894, 0.9663865546218487]\n",
        "\n",
        "# List of the names of the datasets \n",
        "labels = ['bottle', 'cable', 'capsule', 'carpet', 'grid', 'leather', 'metal_nut', 'pill', 'screw', 'tile', 'toothbrush', 'transistor', 'wood', 'zipper']\n",
        "\n",
        "# Average score per list \n",
        "average_image_roc_auc_1 = sum(results_fcor_0_1)/len(results_fcor_0_1)\n",
        "average_image_roc_auc_2 = sum(results_fcor_0_25)/len(results_fcor_0_25)\n",
        "print(\"Average score over all the dataset tested with f_coreset = 0.1: \", round(average_image_roc_auc_1, 3))\n",
        "print(\"Average score over all the datasets tested with f_coreset = 0.25:\", round(average_image_roc_auc_2, 3))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PJ41JWJj-0Fi",
        "outputId": "fb9c2eef-36fd-465c-dc40-7330cc530842"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Average score over all the dataset tested with f_coreset = 0.1:  0.971\n",
            "Average score over all the datasets tested with f_coreset = 0.25: 0.968\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "-- **VISUALIZATION** -- "
      ],
      "metadata": {
        "id": "mgETTEljq9S2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import seaborn as sns\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "data = {'ROC AUC': results_fcor_0_1, 'Dataset': labels}\n",
        "df = pd.DataFrame(data)\n",
        "df = df.sort_values('ROC AUC', ascending = False)\n",
        "plt.figure(figsize=(10,5))\n",
        "plt.xticks(rotation=90)\n",
        "plt.ylim(0.8, 1.0)\n",
        "sns.barplot(x = 'Dataset', y = 'ROC AUC', data = df)\n",
        "plt.title('ROC AUC Scores per Dataset with f_coreset = 0.1')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "id": "2WuAT70UAYRK",
        "outputId": "f9dfc66e-bbb8-47dd-cb67-6abe7c66a832"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data = {'ROC AUC': results_fcor_0_25, 'Dataset': labels}\n",
        "df = pd.DataFrame(data)\n",
        "df = df.sort_values('ROC AUC', ascending = False)\n",
        "plt.figure(figsize=(10,5))\n",
        "plt.xticks(rotation=90)\n",
        "plt.ylim(0.8, 1.0)\n",
        "sns.barplot(x = 'Dataset', y = 'ROC AUC', data = df)\n",
        "plt.title('ROC AUC Scores per Dataset with f_coreset_0_25')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "id": "LbvP0PvxcTGr",
        "outputId": "ce05ade6-a333-484d-c86c-4b5a989639b7"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df = pd.DataFrame({'results_fcor_0_1' : results_fcor_0_1, 'results_fcor_0_25' : results_fcor_0_25, 'labels' : labels})\n",
        "plt.figure(figsize=(10,5))\n",
        "plt.xticks(rotation=90)\n",
        "sns.lineplot(x = 'labels', y = 'value', hue = 'variable', data = pd.melt(df, ['labels']))\n",
        "plt.legend(title = 'Score ROC_AUC comparison')\n",
        "plt.xlabel('dataset name')\n",
        "plt.ylabel('roc_auc score')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        },
        "id": "4V5mm5vkp2Dt",
        "outputId": "00e63738-6006-4fc2-96f0-1f55e6ead37a"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}