{
  "id": 346286,
  "title": "Private Dataset evaluation",
  "url": "/competitions/amex-default-prediction/discussion/346286",
  "author_name": "",
  "post_date": "2022-08-18T18:43:22.123300800Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>I am curious as to how are we going to be evaluated on the private dataset because most of us have transformed data especially the ones who transformed the data into feather and parquet files?</p>",
  "messages": [
    {
      "id": "1905132",
      "postDate": "08/18/2022 18:43:22",
      "content": "<p>I am curious as to how are we going to be evaluated on the private dataset because most of us have transformed data especially the ones who transformed the data into feather and parquet files?</p>",
      "rawMarkdown": "I am curious as to how are we going to be evaluated on the private dataset because most of us have transformed data especially the ones who transformed the data into feather and parquet files?",
      "votes": null
    },
    {
      "id": "1905170",
      "postDate": "08/18/2022 19:21:10",
      "content": "<p>You submission file will be evaluated on the remainder of the data not used in the public leaderboard. This has no relation to the training data (parquet/ feather files)…<br>\nI hope this clarifies this question..</p>",
      "rawMarkdown": "You submission file will be evaluated on the remainder of the data not used in the public leaderboard. This has no relation to the training data (parquet/ feather files)...\nI hope this clarifies this question..",
      "votes": null
    },
    {
      "id": "1905199",
      "postDate": "08/18/2022 19:52:13",
      "content": "<p>I meant<br>\nSuppose my notebook uses the converted data(parquet format) so that means to run it on private dataset, I'd need it to myself to convert it to parquet first<br>\nHow would that happen?</p>",
      "rawMarkdown": "I meant\nSuppose my notebook uses the converted data(parquet format) so that means to run it on private dataset, I'd need it to myself to convert it to parquet first\nHow would that happen?",
      "votes": null
    },
    {
      "id": "1905226",
      "postDate": "08/18/2022 20:41:14",
      "content": "<blockquote>\n  <p>so that means to run it on private dataset</p>\n</blockquote>\n<p>in this competition we submit predictions, not model weights, so there is no need to make inference again  - \"run it on private dataset\" - as in code competitions)</p>\n<p>the .csv file we submit contains both public/private dataset customers/predictions - we only see the score for Public part now, after competition end we'll see the Private scores as well</p>",
      "rawMarkdown": ">so that means to run it on private dataset\n\nin this competition we submit predictions, not model weights, so there is no need to make inference again  - \"run it on private dataset\" - as in code competitions)\n\nthe .csv file we submit contains both public/private dataset customers/predictions - we only see the score for Public part now, after competition end we'll see the Private scores as well",
      "votes": null
    },
    {
      "id": "1905329",
      "postDate": "08/19/2022 00:47:42",
      "content": "<p>I get it now, thanks buddy </p>",
      "rawMarkdown": "I get it now, thanks buddy",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1905170,
      "author_name": "ravi20076",
      "author_url": "",
      "post_date": "08/18/2022 19:21:10",
      "content": "<p>You submission file will be evaluated on the remainder of the data not used in the public leaderboard. This has no relation to the training data (parquet/ feather files)…<br>\nI hope this clarifies this question..</p>",
      "votes": null,
      "replies": [
        {
          "id": 1905199,
          "author_name": "bhavesjain",
          "author_url": "",
          "post_date": "08/18/2022 19:52:13",
          "content": "<p>I meant<br>\nSuppose my notebook uses the converted data(parquet format) so that means to run it on private dataset, I'd need it to myself to convert it to parquet first<br>\nHow would that happen?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1905226,
          "author_name": "imeintanis",
          "author_url": "",
          "post_date": "08/18/2022 20:41:14",
          "content": "<blockquote>\n  <p>so that means to run it on private dataset</p>\n</blockquote>\n<p>in this competition we submit predictions, not model weights, so there is no need to make inference again  - \"run it on private dataset\" - as in code competitions)</p>\n<p>the .csv file we submit contains both public/private dataset customers/predictions - we only see the score for Public part now, after competition end we'll see the Private scores as well</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1905329,
          "author_name": "bhavesjain",
          "author_url": "",
          "post_date": "08/19/2022 00:47:42",
          "content": "<p>I get it now, thanks buddy </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1905132": "I am curious as to how are we going to be evaluated on the private dataset because most of us have transformed data especially the ones who transformed the data into feather and parquet files?",
    "1905170": "You submission file will be evaluated on the remainder of the data not used in the public leaderboard. This has no relation to the training data (parquet/ feather files)...\nI hope this clarifies this question..",
    "1905199": "I meant\nSuppose my notebook uses the converted data(parquet format) so that means to run it on private dataset, I'd need it to myself to convert it to parquet first\nHow would that happen?",
    "1905226": ">so that means to run it on private dataset\n\nin this competition we submit predictions, not model weights, so there is no need to make inference again  - \"run it on private dataset\" - as in code competitions)\n\nthe .csv file we submit contains both public/private dataset customers/predictions - we only see the score for Public part now, after competition end we'll see the Private scores as well",
    "1905329": "I get it now, thanks buddy"
  },
  "source": "meta"
}