{
  "id": 52786,
  "title": "Old test data survives through blends",
  "url": "/competitions/talkingdata-adtracking-fraud-detection/discussion/52786",
  "author_name": "",
  "post_date": "2018-03-23T05:55:07.413327600Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>As the test size was reduced. There are still kernels with 500 MB output data and forked scripts are taking a while for submission. Doesn't this slow down the submission process for others as submissions are evaluated in FIFO? Kindly remove the old tests as there are a bunch of blends floating around. The old test data might never RIP. </p>",
  "messages": [
    {
      "id": "301697",
      "postDate": "03/23/2018 05:55:07",
      "content": "<p>As the test size was reduced. There are still kernels with 500 MB output data and forked scripts are taking a while for submission. Doesn't this slow down the submission process for others as submissions are evaluated in FIFO? Kindly remove the old tests as there are a bunch of blends floating around. The old test data might never RIP. </p>",
      "rawMarkdown": "As the test size was reduced. There are still kernels with 500 MB output data and forked scripts are taking a while for submission. Doesn't this slow down the submission process for others as submissions are evaluated in FIFO? Kindly remove the old tests as there are a bunch of blends floating around. The old test data might never RIP.",
      "votes": null
    },
    {
      "id": "301762",
      "postDate": "03/23/2018 07:52:18",
      "content": "<p>The test size was reduced?!? Could you please share a link to the source you get this information from. All I know that the \"old\" / initial test set was publicly shared once again, but it is only for feature engineering and is actually larger than the one for evaluation not the other way round.</p>",
      "rawMarkdown": "The test size was reduced?!? Could you please share a link to the source you get this information from. All I know that the \"old\" / initial test set was publicly shared once again, but it is only for feature engineering and is actually larger than the one for evaluation not the other way round.",
      "votes": null
    },
    {
      "id": "301798",
      "postDate": "03/23/2018 09:08:30",
      "content": "<p>Please do check this - \n<a href=\"https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51210\">https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51210</a></p>",
      "rawMarkdown": "Please do check this - \nhttps://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51210",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 301762,
      "author_name": "asparuhhristov",
      "author_url": "",
      "post_date": "03/23/2018 07:52:18",
      "content": "<p>The test size was reduced?!? Could you please share a link to the source you get this information from. All I know that the \"old\" / initial test set was publicly shared once again, but it is only for feature engineering and is actually larger than the one for evaluation not the other way round.</p>",
      "votes": null,
      "replies": [
        {
          "id": 301798,
          "author_name": "shaz13",
          "author_url": "",
          "post_date": "03/23/2018 09:08:30",
          "content": "<p>Please do check this - \n<a href=\"https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51210\">https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51210</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "301697": "As the test size was reduced. There are still kernels with 500 MB output data and forked scripts are taking a while for submission. Doesn't this slow down the submission process for others as submissions are evaluated in FIFO? Kindly remove the old tests as there are a bunch of blends floating around. The old test data might never RIP.",
    "301762": "The test size was reduced?!? Could you please share a link to the source you get this information from. All I know that the \"old\" / initial test set was publicly shared once again, but it is only for feature engineering and is actually larger than the one for evaluation not the other way round.",
    "301798": "Please do check this - \nhttps://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51210"
  },
  "source": "meta"
}