{
  "id": 153382,
  "title": "How to annotate the public test images?",
  "url": "/competitions/global-wheat-detection/discussion/153382",
  "author_name": "",
  "post_date": "2020-05-24T13:40:06.183921400Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi,\nI'm new to kaggle, so I might be missing the point, but how can\\should I annotate the 10 test images and why aren't they annotated in the first place?</p>",
  "messages": [
    {
      "id": "859495",
      "postDate": "05/24/2020 13:40:06",
      "content": "<p>Hi,\nI'm new to kaggle, so I might be missing the point, but how can\\should I annotate the 10 test images and why aren't they annotated in the first place?</p>",
      "rawMarkdown": "Hi,\nI'm new to kaggle, so I might be missing the point, but how can\\should I annotate the 10 test images and why aren't they annotated in the first place?",
      "votes": null
    },
    {
      "id": "859611",
      "postDate": "05/24/2020 15:26:24",
      "content": "<p>Hi, \nThe point of this competition is to train a model that will be able to accurately find the wheat heads in the test images. The test images are not to be annotated manually. You have to train your model on the given data and then predict what the annotations should be for the test data.</p>",
      "rawMarkdown": "Hi, \nThe point of this competition is to train a model that will be able to accurately find the wheat heads in the test images. The test images are not to be annotated manually. You have to train your model on the given data and then predict what the annotations should be for the test data.",
      "votes": null
    },
    {
      "id": "859627",
      "postDate": "05/24/2020 15:41:56",
      "content": "<p>The prediction for the competition are on a private held-out test set. Or am I mistaken?\nWhy not give us annotation on the public (10 images) test set?</p>",
      "rawMarkdown": "The prediction for the competition are on a private held-out test set. Or am I mistaken?\nWhy not give us annotation on the public (10 images) test set?",
      "votes": null
    },
    {
      "id": "859814",
      "postDate": "05/24/2020 19:03:45",
      "content": "<p>Although the test set is private, I think the 10 images are a part of it so it's not logical to give the labels of them. If your point is to check your predictions on the 10 images, you can plot them with the predicted bboxes. And if your point is to check the performance of your model you can still split you training data and use a part of it for the test. </p>",
      "rawMarkdown": "Although the test set is private, I think the 10 images are a part of it so it's not logical to give the labels of them. If your point is to check your predictions on the 10 images, you can plot them with the predicted bboxes. And if your point is to check the performance of your model you can still split you training data and use a part of it for the test.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 859611,
      "author_name": "amoghjrules",
      "author_url": "",
      "post_date": "05/24/2020 15:26:24",
      "content": "<p>Hi, \nThe point of this competition is to train a model that will be able to accurately find the wheat heads in the test images. The test images are not to be annotated manually. You have to train your model on the given data and then predict what the annotations should be for the test data.</p>",
      "votes": null,
      "replies": [
        {
          "id": 859627,
          "author_name": "adizhol",
          "author_url": "",
          "post_date": "05/24/2020 15:41:56",
          "content": "<p>The prediction for the competition are on a private held-out test set. Or am I mistaken?\nWhy not give us annotation on the public (10 images) test set?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 859814,
          "author_name": "ffares",
          "author_url": "",
          "post_date": "05/24/2020 19:03:45",
          "content": "<p>Although the test set is private, I think the 10 images are a part of it so it's not logical to give the labels of them. If your point is to check your predictions on the 10 images, you can plot them with the predicted bboxes. And if your point is to check the performance of your model you can still split you training data and use a part of it for the test. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "859495": "Hi,\nI'm new to kaggle, so I might be missing the point, but how can\\should I annotate the 10 test images and why aren't they annotated in the first place?",
    "859611": "Hi, \nThe point of this competition is to train a model that will be able to accurately find the wheat heads in the test images. The test images are not to be annotated manually. You have to train your model on the given data and then predict what the annotations should be for the test data.",
    "859627": "The prediction for the competition are on a private held-out test set. Or am I mistaken?\nWhy not give us annotation on the public (10 images) test set?",
    "859814": "Although the test set is private, I think the 10 images are a part of it so it's not logical to give the labels of them. If your point is to check your predictions on the 10 images, you can plot them with the predicted bboxes. And if your point is to check the performance of your model you can still split you training data and use a part of it for the test."
  },
  "source": "meta"
}