{
  "id": 66023,
  "title": "Local Test data independent of Training data?",
  "url": "/competitions/airbus-ship-detection/discussion/66023",
  "author_name": "",
  "post_date": "2018-09-17T12:53:10.276154300Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi, I understand that most of the ships in training images are found twice. Is there a simple way to separate a local test set of images from the training set, and be sure that the examples in this local test set are not duplicated in the training set (which would spoil the local test)?</p>",
  "messages": [
    {
      "id": "388701",
      "postDate": "09/17/2018 12:53:10",
      "content": "<p>Hi, I understand that most of the ships in training images are found twice. Is there a simple way to separate a local test set of images from the training set, and be sure that the examples in this local test set are not duplicated in the training set (which would spoil the local test)?</p>",
      "rawMarkdown": "Hi, I understand that most of the ships in training images are found twice. Is there a simple way to separate a local test set of images from the training set, and be sure that the examples in this local test set are not duplicated in the training set (which would spoil the local test)?",
      "votes": null
    },
    {
      "id": "388747",
      "postDate": "09/17/2018 13:56:08",
      "content": "<p>It has been discussed:</p>\n\n<ul>\n<li>It would be great if Competition Host provides overlap information. <a href=\"https://www.kaggle.com/c/airbus-ship-detection/discussion/64675\">https://www.kaggle.com/c/airbus-ship-detection/discussion/64675</a></li>\n<li>Calculate the overlap yourself <a href=\"https://www.kaggle.com/c/airbus-ship-detection/discussion/64588\">https://www.kaggle.com/c/airbus-ship-detection/discussion/64588</a></li>\n</ul>",
      "rawMarkdown": "It has been discussed:\n\n - It would be great if Competition Host provides overlap information. https://www.kaggle.com/c/airbus-ship-detection/discussion/64675\n - Calculate the overlap yourself [https://www.kaggle.com/c/airbus-ship-detection/discussion/64588][1]\n\n\n  [1]: https://www.kaggle.com/c/airbus-ship-detection/discussion/64588",
      "votes": null
    },
    {
      "id": "388761",
      "postDate": "09/17/2018 14:24:38",
      "content": "<p>I am aware of these discussions. As more than 130 teams have now made the 1.00 score, I was just hoping for some kind of 'Ready Made' solution... </p>",
      "rawMarkdown": "I am aware of these discussions. As more than 130 teams have now made the 1.00 score, I was just hoping for some kind of 'Ready Made' solution...",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 388747,
      "author_name": "waskita",
      "author_url": "",
      "post_date": "09/17/2018 13:56:08",
      "content": "<p>It has been discussed:</p>\n\n<ul>\n<li>It would be great if Competition Host provides overlap information. <a href=\"https://www.kaggle.com/c/airbus-ship-detection/discussion/64675\">https://www.kaggle.com/c/airbus-ship-detection/discussion/64675</a></li>\n<li>Calculate the overlap yourself <a href=\"https://www.kaggle.com/c/airbus-ship-detection/discussion/64588\">https://www.kaggle.com/c/airbus-ship-detection/discussion/64588</a></li>\n</ul>",
      "votes": null,
      "replies": []
    },
    {
      "id": 388761,
      "author_name": "petersorensen360",
      "author_url": "",
      "post_date": "09/17/2018 14:24:38",
      "content": "<p>I am aware of these discussions. As more than 130 teams have now made the 1.00 score, I was just hoping for some kind of 'Ready Made' solution... </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "388701": "Hi, I understand that most of the ships in training images are found twice. Is there a simple way to separate a local test set of images from the training set, and be sure that the examples in this local test set are not duplicated in the training set (which would spoil the local test)?",
    "388747": "It has been discussed:\n\n - It would be great if Competition Host provides overlap information. https://www.kaggle.com/c/airbus-ship-detection/discussion/64675\n - Calculate the overlap yourself [https://www.kaggle.com/c/airbus-ship-detection/discussion/64588][1]\n\n\n  [1]: https://www.kaggle.com/c/airbus-ship-detection/discussion/64588",
    "388761": "I am aware of these discussions. As more than 130 teams have now made the 1.00 score, I was just hoping for some kind of 'Ready Made' solution..."
  },
  "source": "meta"
}