imlmodontop.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.svm import LinearSVC, SVC\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20828, 128) (20828,) (5951, 128) (5951,) (2976, 128) (2976,)\n",
      "Linear SVM val score: 0.8902705427659217\n",
      "Linear SVM test score: 0.8877688172043011\n",
      "KNN: 0.891950932616367\n",
      "KNN test score: 0.8951612903225806\n",
      "SVM: 0.8932952444967233\n",
      "SVM test score: 0.8924731182795699\n",
      "RF: 0.8922870105864561\n",
      "RF test score: 0.8938172043010753\n"
     ]
    }
   ],
   "source": [
    "\n",
    "if __name__ == '__main__':\n",
    "\n",
    "    root = \"/mnt/sda/nikhil/pbb/\"\n",
    "\n",
    "    x_train, y_train = np.load(root + 'imlstagetrain_embeddings.npy'), np.load(root + 'imlstagetrain_labels.npy')\n",
    "    x_val, y_val = np.load(root + 'imlstagevalid_embeddings.npy'), np.load(root + 'imlstagevalid_labels.npy')\n",
    "    x_test, y_test = np.load(root + 'imlstagetest_embeddings.npy'), np.load(root + 'imlstagetest_labels.npy')\n",
    "\n",
    "    print(x_train.shape, y_train.shape, x_val.shape, y_val.shape, x_test.shape, y_test.shape)\n",
    "\n",
    "    L_SVM = LinearSVC()\n",
    "    L_SVM.fit(x_train, y_train)\n",
    "    val_score = L_SVM.score(x_val, y_val)\n",
    "    print('Linear SVM val score:', val_score)\n",
    "    test_score = L_SVM.score(x_test, y_test)\n",
    "    print('Linear SVM test score:', test_score)\n",
    "\n",
    "    knn = KNeighborsClassifier()\n",
    "    knn.fit(x_train, y_train)\n",
    "    score = knn.score(x_val, y_val)\n",
    "    print('KNN:', score)\n",
    "    test_score = knn.score(x_test, y_test)\n",
    "    print('KNN test score:', test_score)\n",
    "\n",
    "    SVM = SVC()\n",
    "    SVM.fit(x_train, y_train)\n",
    "    score = SVM.score(x_val, y_val)\n",
    "    print('SVM:', score)\n",
    "    test_score = SVM.score(x_test, y_test)\n",
    "    print('SVM test score:', test_score)\n",
    "\n",
    "    RF = RandomForestClassifier()\n",
    "    RF.fit(x_train, y_train)\n",
    "    score = RF.score(x_val, y_val)\n",
    "    print('RF:', score)\n",
    "    test_score = RF.score(x_test, y_test)\n",
    "    print('RF test score:', test_score)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import os\n",
    "# import pandas as pd\n",
    "\n",
    "# test_path = \"/mnt/sda/nikhil/pbb/iml/test/\"\n",
    "\n",
    "# file_names = []\n",
    "# ground_truth = []\n",
    "\n",
    "# #Recursively search for image files in all subfolders within test_path\n",
    "# for root, dirs, files in os.walk(test_path):\n",
    "#     for file_name in files:\n",
    "#         file_ext = os.path.splitext(file_name)[1].lower()\n",
    "#         if file_ext in ['.jpg', '.jpeg', '.png', '.gif']:\n",
    "#             file_names.append(file_name)\n",
    "#             label = os.path.basename(root)\n",
    "#             if label == 'init':\n",
    "#                 ground_truth.append('황적갈')\n",
    "#             elif label == 'mid':\n",
    "#                 ground_truth.append('회백색')\n",
    "#             elif label == 'late':\n",
    "#                 ground_truth.append('탈리')\n",
    "\n",
    "# # Make predictions for the test data\n",
    "# test_predictions = knn.predict(x_test)\n",
    "\n",
    "# # Create a DataFrame with filenames, ground truth, and predicted labels\n",
    "# test_df = pd.DataFrame({'Filename': file_names, 'Ground Truth': ground_truth, 'Predicted Label': test_predictions})\n",
    "\n",
    "# # Save the DataFrame to a CSV file\n",
    "# test_df.to_csv('imlstagetest_predictions.csv', index=False)\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "\n",
    "test_path = \"/mnt/sda/nikhil/pbb/iml/test/\"\n",
    "\n",
    "file_names = []\n",
    "ground_truth = []\n",
    "\n",
    "# Recursively search for image files in all subfolders within test_path\n",
    "for root, dirs, files in os.walk(test_path):\n",
    "    for file_name in files:\n",
    "        file_ext = os.path.splitext(file_name)[1].lower()\n",
    "        if file_ext in ['.jpg', '.jpeg', '.png', '.gif']:\n",
    "            file_names.append(file_name)\n",
    "            ground_truth.append(os.path.basename(root))\n",
    "\n",
    "# Make predictions for the test data\n",
    "test_predictions = knn.predict(x_test)\n",
    "\n",
    "# Create a DataFrame with filenames, ground truth, and predicted labels\n",
    "test_df = pd.DataFrame({'Filename': file_names, 'Ground Truth': ground_truth, 'Predicted Label': test_predictions})\n",
    "\n",
    "# Save the DataFrame to a CSV file\n",
    "test_df.to_csv('imlstagetest_predictions.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "# Compute the confusion matrix\n",
    "cm = confusion_matrix(test_df['Ground Truth'], test_df['Predicted Label'])\n",
    "\n",
    "# Get class labels from unique values in Ground Truth\n",
    "labels = test_df['Ground Truth'].unique()\n",
    "\n",
    "# Plot the confusion matrix with annotations and values\n",
    "plt.figure(figsize=(8, 6))\n",
    "sns.heatmap(cm, annot=True, fmt=\"d\", cmap=\"Blues\", xticklabels=labels, yticklabels=labels)\n",
    "plt.xlabel('Predicted Labels')\n",
    "plt.ylabel('Ground Truth Labels')\n",
    "plt.title('Confusion Matrix')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "# Load the existing CSV file\n",
    "test_df = pd.read_csv('test_predictions.csv')\n",
    "\n",
    "# Create combined ground truth and combined label columns\n",
    "test_df['Combined Ground Truth'] = test_df['Ground Truth']\n",
    "test_df['Combined Label'] = test_df['Predicted Label']\n",
    "\n",
    "# Replace 'pine' and 'black' with 'pineblack' in the combined columns\n",
    "test_df.loc[test_df['Combined Ground Truth'].isin(['pine', 'black']), 'Combined Ground Truth'] = 'pineblack'\n",
    "test_df.loc[test_df['Combined Label'].isin(['pine', 'black']), 'Combined Label'] = 'pineblack'\n",
    "\n",
    "# Save the modified DataFrame back to the CSV file\n",
    "test_df.to_csv('test_predictions.csv', index=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[ 777    6   36]\n",
      " [  10  259   47]\n",
      " [  55   50 1537]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "# Calculate the confusion matrix\n",
    "cm = confusion_matrix(test_df['Ground Truth'], test_df['Predicted Label'])\n",
    "\n",
    "# Display the confusion matrix\n",
    "print(\"Confusion Matrix:\")\n",
    "print(cm)\n",
    "\n",
    "cm = confusion_matrix(test_df['Ground Truth'], test_df['Predicted Label'])\n",
    "\n",
    "# Plot the confusion matrix\n",
    "plt.figure(figsize=(8, 6))\n",
    "sns.heatmap(cm, annot=True, fmt='d', cmap='Blues')\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Labels')\n",
    "plt.ylabel('Ground Truth Labels')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[ 777    6   36]\n",
      " [  10  259   47]\n",
      " [  55   50 1537]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Calculate the confusion matrix\n",
    "cm = confusion_matrix(test_df['Ground Truth'], test_df['Predicted Label'])\n",
    "\n",
    "# Display the confusion matrix\n",
    "print(\"Confusion Matrix:\")\n",
    "print(cm)\n",
    "\n",
    "# Plot the confusion matrix\n",
    "plt.figure(figsize=(8, 6))\n",
    "sns.heatmap(cm, annot=True, fmt='d', cmap='Blues')\n",
    "\n",
    "# Annotate the labels\n",
    "tick_labels = ['Beauty', 'Black', 'Pine']\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Labels')\n",
    "plt.ylabel('Ground Truth Labels')\n",
    "plt.xticks(ticks=[0, 1, 2], labels=tick_labels)\n",
    "plt.yticks(ticks=[0, 1, 2], labels=tick_labels)\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(19414, 128) (19414,) (5546, 128) (5546,) (2777, 128) (2777,)\n",
      "Linear SVM validation score: 0.923007573025604\n",
      "Linear SVM test score: 0.9265394310406914\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_1258746/3320569483.py:20: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  val_predictions_combined = np.where(np.logical_or(val_predictions == 1, val_predictions == 2), 1, val_predictions)\n",
      "/tmp/ipykernel_1258746/3320569483.py:30: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  test_predictions_combined = np.where(np.logical_or(test_predictions == 1, test_predictions == 2), 1, test_predictions)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.svm import LinearSVC\n",
    "import numpy as np\n",
    "\n",
    "if __name__ == '__main__':\n",
    "    root = '/mnt/sda/nikhil/pbb/'\n",
    "\n",
    "    x_train, y_train = np.load(root + 'train_embeddings.npy'), np.load(root + 'train_labels.npy')\n",
    "    x_val, y_val = np.load(root + 'valid_embeddings.npy'), np.load(root + 'valid_labels.npy')\n",
    "    x_test, y_test = np.load(root + 'test_embeddings.npy'), np.load(root + 'test_labels.npy')\n",
    "\n",
    "    print(x_train.shape, y_train.shape, x_val.shape, y_val.shape, x_test.shape, y_test.shape)\n",
    "\n",
    "    L_SVM = LinearSVC()\n",
    "    L_SVM.fit(x_train, y_train)\n",
    "\n",
    "    # Make predictions for validation set\n",
    "    val_predictions = L_SVM.predict(x_val)\n",
    "\n",
    "    # Combine classes 1 and 2 into a single class 1\n",
    "    val_predictions_combined = np.where(np.logical_or(val_predictions == 1, val_predictions == 2), 1, val_predictions)\n",
    "\n",
    "    # Calculate validation score\n",
    "    val_score = np.mean(val_predictions_combined == y_val)\n",
    "    print('Linear SVM validation score:', val_score)\n",
    "\n",
    "    # Make predictions for test set\n",
    "    test_predictions = L_SVM.predict(x_test)\n",
    "\n",
    "    # Combine classes 1 and 2 into a single class 1\n",
    "    test_predictions_combined = np.where(np.logical_or(test_predictions == 1, test_predictions == 2), 1, test_predictions)\n",
    "\n",
    "    # Calculate test score\n",
    "    test_score = np.mean(test_predictions_combined == y_test)\n",
    "    print('Linear SVM test score:', test_score)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[1893   65]\n",
      " [  31  524]]\n",
      "Accuracy: 0.9614160700079555\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score\n",
    "\n",
    "# Read the CSV file into a DataFrame\n",
    "df = pd.read_csv(\"/mnt/sda/nikhil/pbb/test_predictions_final.csv\")\n",
    "\n",
    "# Extract ground truth labels (4th column) and predicted labels (5th column)\n",
    "ground_truth = df.iloc[:, 3].astype(str)\n",
    "predictions = df.iloc[:, 4].astype(str)\n",
    "\n",
    "# Define the class labels\n",
    "labels = ['pineblack', 'beauty']\n",
    "\n",
    "# Calculate the confusion matrix\n",
    "cm = confusion_matrix(ground_truth, predictions, labels=labels)\n",
    "\n",
    "# Calculate the accuracy\n",
    "accuracy = accuracy_score(ground_truth, predictions)\n",
    "\n",
    "# Display the confusion matrix and accuracy\n",
    "print(\"Confusion Matrix:\")\n",
    "print(cm)\n",
    "print(\"Accuracy:\", accuracy)\n",
    "\n",
    "# Plot the confusion matrix\n",
    "plt.figure(figsize=(8, 6))\n",
    "sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', xticklabels=labels, yticklabels=labels)\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Labels')\n",
    "plt.ylabel('Ground Truth Labels')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of image files:  29755\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "test_path = \"/mnt/sda/nikhil/pbb/img/\"\n",
    "image_extensions = ['.jpg', '.jpeg', '.png']\n",
    "\n",
    "def count_image_files(path):\n",
    "    count = 0\n",
    "    for root, dirs, files in os.walk(path):\n",
    "        for file in files:\n",
    "            if any(file.lower().endswith(ext) for ext in image_extensions):\n",
    "                count += 1\n",
    "    return count\n",
    "\n",
    "image_count = count_image_files(test_path)\n",
    "print(\"Number of image files: \", image_count)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ptc",
   "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": "3.9.16"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}