{
  "id": 338863,
  "title": "Have you trained on any external dataset yet?",
  "url": "/competitions/AI4Code/discussion/338863",
  "author_name": "",
  "post_date": "2022-07-22T10:02:05.063804500Z",
  "votes": 7,
  "comment_count": 3,
  "views": 0,
  "content": "<p>For me I used the Kaggle notebooks dataset shared <a href=\"https://www.kaggle.com/competitions/AI4Code/discussion/326520\" target=\"_blank\">here</a>, and it gives ~0.01 boost on the LB (maybe the leak)</p>\n<p>Just wandering if the people in the front of the leaderboard trained on any external dataset yet or not…</p>",
  "messages": [
    {
      "id": "1866191",
      "postDate": "07/22/2022 10:02:05",
      "content": "<p>For me I used the Kaggle notebooks dataset shared <a href=\"https://www.kaggle.com/competitions/AI4Code/discussion/326520\" target=\"_blank\">here</a>, and it gives ~0.01 boost on the LB (maybe the leak)</p>\n<p>Just wandering if the people in the front of the leaderboard trained on any external dataset yet or not…</p>",
      "rawMarkdown": "For me I used the Kaggle notebooks dataset shared [here](https://www.kaggle.com/competitions/AI4Code/discussion/326520), and it gives ~0.01 boost on the LB (maybe the leak)\n\nJust wandering if the people in the front of the leaderboard trained on any external dataset yet or not...",
      "votes": null
    },
    {
      "id": "1867182",
      "postDate": "07/23/2022 03:52:41",
      "content": "<p>We haven't, we have been thinking about the license nuances/issues if we use external data and need to be careful about the same </p>",
      "rawMarkdown": "We haven't, we have been thinking about the license nuances/issues if we use external data and need to be careful about the same",
      "votes": null
    },
    {
      "id": "1867351",
      "postDate": "07/23/2022 06:42:55",
      "content": "<p>Thanks, good to know that.</p>",
      "rawMarkdown": "Thanks, good to know that.",
      "votes": null
    },
    {
      "id": "1869137",
      "postDate": "07/24/2022 14:15:30",
      "content": "<p>Have not used but planed to use it, I think we could get better PB result using it (also on future test set).</p>",
      "rawMarkdown": "Have not used but planed to use it, I think we could get better PB result using it (also on future test set).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1867182,
      "author_name": "init27",
      "author_url": "",
      "post_date": "07/23/2022 03:52:41",
      "content": "<p>We haven't, we have been thinking about the license nuances/issues if we use external data and need to be careful about the same </p>",
      "votes": null,
      "replies": [
        {
          "id": 1867351,
          "author_name": "snaker",
          "author_url": "",
          "post_date": "07/23/2022 06:42:55",
          "content": "<p>Thanks, good to know that.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1869137,
      "author_name": "goldenlock",
      "author_url": "",
      "post_date": "07/24/2022 14:15:30",
      "content": "<p>Have not used but planed to use it, I think we could get better PB result using it (also on future test set).</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1866191": "For me I used the Kaggle notebooks dataset shared [here](https://www.kaggle.com/competitions/AI4Code/discussion/326520), and it gives ~0.01 boost on the LB (maybe the leak)\n\nJust wandering if the people in the front of the leaderboard trained on any external dataset yet or not...",
    "1867182": "We haven't, we have been thinking about the license nuances/issues if we use external data and need to be careful about the same",
    "1867351": "Thanks, good to know that.",
    "1869137": "Have not used but planed to use it, I think we could get better PB result using it (also on future test set)."
  },
  "source": "meta"
}