{
  "fetched_at": "2026-10-19T09:51:15Z",
  "source_url": "https://www.kaggle.com/competitions/car-classificationproject-vision/data",
  "pages": [
    {
      "name": "data-description",
      "content": "<h1>Overview</h1>\n<p> The data has been split into two groups:</p>\n<ul>\n<li>training set(train folder)</li>\n<li>test set(test folder)</li>\n<p>The training set contains 45 different classes(cars) and there are 100 images for each car. These images are stored in separate folders according to names of the cars. The labels of the images are the folder names.</p>\n<p>The test set contains 450 pictures of different cars.</p>\n<p>Train your model on the training set. Then use that model to generate predictions and store them in a csv file</p>\n<p> The csv file must contain two columns. The first column will have the string \"image\"+image_no. The second column will have the predictions in the form of integers(0-44). You have to assign an integer to each car when the car names are sorted without considering the case</p>\n<p> If you have any problem regarding the label encoding of the car names, refer car_map.csv</p>\n<p> If you have any problem regarding the format of the file to be submitted, refer sample_submission.csv</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "<p>Cars are one of the most sought after luxury today. They are mostly admired for their sleek design and features.</p>\n<p>In this task, you are given 45 classes(cars) with 100 images each.Train a CNN on the given dataset and submit the predictions on the test set in a csv file. Download the dataset from data tab and read further instructions. </p>\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "<h1>Public-Private Leaderboard</h1>\n<p>\nThe test set has been further split into a private and a public set on a 30-70 basis. On submission, the score as seen on the public leaderboard is based on the performance set on the public test set, the final scores post deadline will be based on the performance on the private test set. This split has been adopted in order to prevent any participants from overfitting on the public leaderboard. \n</p>\n<hr>\n<h1>Evaluation Metric</h1>\n<p>The evaluation metric for this competition is Categorization Accuracy i.e. the percentage of correct predictions on total submissions. \nThis will be given by,</p>\n<h2>$$x = {t \\over n}\\cdot 100$$</h2>\n<p>where t is the number of correct classification and n is the total number of samples</p>\n<hr>\n<h1>Submission Format</h1>\n<p>You must predict the class no. for each <b><i>image_id</i></b>. Here, the predicted class must belong to one of the 45 classes of cars as mentioned in the training set. Here the class number is mapped to the classes of cars as mentioned in the car_map.csv file.\nThe submission is to be made in a csv file format, must have a header and should look like the following:</p>\n<pre>ImageId,Class<br>image1.jpg,42<br></pre>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "Task Guidelines:\n\n - Only one account per participant.\n - Submission Limits: a maximum of 20 submissions per day.\n - On completion of the task add a well documented notebook pertaining to the task under the notebooks section, make the \n    notebook public.\n    \n**All the Best!!**\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}