{
  "id": 285366,
  "title": "what is supplemental_train.csv file...",
  "url": "/competitions/g-research-crypto-forecasting/discussion/285366",
  "author_name": "Roman Kazmin",
  "post_date": "2021-11-04T12:02:28.280000",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Please explain what is supplemental_train.csv data</p>",
  "messages": [
    {
      "id": 1571004,
      "postDate": "2021-11-04T14:36:30.870Z",
      "content": "<p>I understand this to mean that supplemental_data.csv is a placeholder file which will be replaced with a new file of the same name, containing new data, roughly every two weeks after submissions are finalized.  Submitted notebooks will use the supplemental_data.csv for inference during the evaluation period.  </p>\n<p>I'm also wondering if notebooks will be able to retrain and pickle updates using the new data - will ask in a new thread.  </p>",
      "rawMarkdown": "I understand this to mean that supplemental_data.csv is a placeholder file which will be replaced with a new file of the same name, containing new data, roughly every two weeks after submissions are finalized.  Submitted notebooks will use the supplemental_data.csv for inference during the evaluation period.  \n\nI'm also wondering if notebooks will be able to retrain and pickle updates using the new data - will ask in a new thread.  ",
      "votes": 3
    },
    {
      "id": 1570823,
      "postDate": "2021-11-04T12:02:28.280Z",
      "content": "<p>Please explain what is supplemental_train.csv data</p>",
      "rawMarkdown": "Please explain what is supplemental_train.csv data",
      "votes": 2
    },
    {
      "id": 1605971,
      "postDate": "2021-12-04T18:46:37.943Z",
      "content": "<p>According to this thread here, supplemental_train is just a subset of the training file. Does not quite answer the exact purpose thought. <a href=\"https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/286829\" target=\"_blank\">https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/286829</a></p>",
      "rawMarkdown": "According to this thread here, supplemental_train is just a subset of the training file. Does not quite answer the exact purpose thought. https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/286829"
    },
    {
      "id": 1570831,
      "postDate": "2021-11-04T12:08:22.743Z",
      "rawMarkdown": "",
      "isDeleted": true,
      "replies": [
        {
          "id": 1570976,
          "postDate": "2021-11-04T14:20:52.097Z",
          "content": "<p>Hi, for supplemental_train.csv, I am still confused about the description. I don't know what the use of this file is. What's the significance of the file. That's why I made the post <a href=\"https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946\" target=\"_blank\">https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946</a>.</p>\n<p>Let me illustrate my confusion to you:</p>\n<blockquote>\n  <p>After the submission period is over this file's data will be replaced with cryptoasset prices from the submission period. </p>\n</blockquote>\n<ol>\n<li><p>What is the significance of replacing the file data with cryptoasset prices from the submission period? Why doing this? What should I do with the file data before the submission period is over?</p></li>\n<li><p>What do you mean by \"cryptoasset prices from the submission period?\" Do you mean that the current data in supplemental_train is old data, then you will replace all of them by new data collected during the training phrase when the training phrase ends? Will these new data like the current one have null values ?</p></li>\n</ol>\n<p>Thanks!</p>",
          "rawMarkdown": "Hi, for supplemental_train.csv, I am still confused about the description. I don't know what the use of this file is. What's the significance of the file. That's why I made the post https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946.\n\nLet me illustrate my confusion to you:\n\n> After the submission period is over this file's data will be replaced with cryptoasset prices from the submission period. \n\n1. What is the significance of replacing the file data with cryptoasset prices from the submission period? Why doing this? What should I do with the file data before the submission period is over?\n\n2. What do you mean by \"cryptoasset prices from the submission period?\" Do you mean that the current data in supplemental_train is old data, then you will replace all of them by new data collected during the training phrase when the training phrase ends? Will these new data like the current one have null values ?\n\nThanks!",
          "votes": 2
        },
        {
          "id": 1596761,
          "postDate": "2021-11-26T20:08:29.410Z",
          "content": "<p>This is a really good question.  Did you every receive clarification of any kind about this question?  Thanks,</p>\n<p>Ed</p>",
          "rawMarkdown": "This is a really good question.  Did you every receive clarification of any kind about this question?  Thanks,\n\nEd"
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 1571004,
      "author_name": "Dave",
      "author_url": "",
      "post_date": "2021-11-04T14:36:30.870000",
      "content": "<p>I understand this to mean that supplemental_data.csv is a placeholder file which will be replaced with a new file of the same name, containing new data, roughly every two weeks after submissions are finalized.  Submitted notebooks will use the supplemental_data.csv for inference during the evaluation period.  </p>\n<p>I'm also wondering if notebooks will be able to retrain and pickle updates using the new data - will ask in a new thread.  </p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 1605971,
      "author_name": "wolfgang",
      "author_url": "",
      "post_date": "2021-12-04T18:46:37.943000",
      "content": "<p>According to this thread here, supplemental_train is just a subset of the training file. Does not quite answer the exact purpose thought. <a href=\"https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/286829\" target=\"_blank\">https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/286829</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1570831,
      "author_name": "",
      "author_url": "",
      "post_date": "2021-11-04T12:08:22.743000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 1570976,
          "author_name": "Yue Sun",
          "author_url": "",
          "post_date": "2021-11-04T14:20:52.097000",
          "content": "<p>Hi, for supplemental_train.csv, I am still confused about the description. I don't know what the use of this file is. What's the significance of the file. That's why I made the post <a href=\"https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946\" target=\"_blank\">https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946</a>.</p>\n<p>Let me illustrate my confusion to you:</p>\n<blockquote>\n  <p>After the submission period is over this file's data will be replaced with cryptoasset prices from the submission period. </p>\n</blockquote>\n<ol>\n<li><p>What is the significance of replacing the file data with cryptoasset prices from the submission period? Why doing this? What should I do with the file data before the submission period is over?</p></li>\n<li><p>What do you mean by \"cryptoasset prices from the submission period?\" Do you mean that the current data in supplemental_train is old data, then you will replace all of them by new data collected during the training phrase when the training phrase ends? Will these new data like the current one have null values ?</p></li>\n</ol>\n<p>Thanks!</p>",
          "votes": 2,
          "replies": []
        },
        {
          "id": 1596761,
          "author_name": "Ed Welch",
          "author_url": "",
          "post_date": "2021-11-26T20:08:29.410000",
          "content": "<p>This is a really good question.  Did you every receive clarification of any kind about this question?  Thanks,</p>\n<p>Ed</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1571004": "I understand this to mean that supplemental_data.csv is a placeholder file which will be replaced with a new file of the same name, containing new data, roughly every two weeks after submissions are finalized.  Submitted notebooks will use the supplemental_data.csv for inference during the evaluation period.  \n\nI'm also wondering if notebooks will be able to retrain and pickle updates using the new data - will ask in a new thread.  ",
    "1570823": "Please explain what is supplemental_train.csv data",
    "1605971": "According to this thread here, supplemental_train is just a subset of the training file. Does not quite answer the exact purpose thought. https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/286829",
    "1570831": ""
  }
}