{
  "id": 313444,
  "title": "Resnet is all you need? ;)",
  "url": "/competitions/ultra-mnist/discussion/313444",
  "author_name": "",
  "post_date": "2022-03-17T06:02:39.178697600Z",
  "votes": 7,
  "comment_count": 5,
  "views": 0,
  "content": "<p>With resent and fastai local validation score is fine (at least for this stage in comp), however as hinted in this discussion (<a href=\"https://www.kaggle.com/c/ultra-mnist/discussion/313437)\" target=\"_blank\">https://www.kaggle.com/c/ultra-mnist/discussion/313437)</a>, lb score is a lot lower than that. probably test set is different than train set?</p>\n<p><a href=\"https://www.kaggle.com/code/miwojc/ultra-mnist-fastai-starter\" target=\"_blank\">https://www.kaggle.com/code/miwojc/ultra-mnist-fastai-starter</a></p>",
  "messages": [
    {
      "id": "1725466",
      "postDate": "03/17/2022 06:02:39",
      "content": "<p>With resent and fastai local validation score is fine (at least for this stage in comp), however as hinted in this discussion (<a href=\"https://www.kaggle.com/c/ultra-mnist/discussion/313437)\" target=\"_blank\">https://www.kaggle.com/c/ultra-mnist/discussion/313437)</a>, lb score is a lot lower than that. probably test set is different than train set?</p>\n<p><a href=\"https://www.kaggle.com/code/miwojc/ultra-mnist-fastai-starter\" target=\"_blank\">https://www.kaggle.com/code/miwojc/ultra-mnist-fastai-starter</a></p>",
      "rawMarkdown": "With resent and fastai local validation score is fine (at least for this stage in comp), however as hinted in this discussion (https://www.kaggle.com/c/ultra-mnist/discussion/313437), lb score is a lot lower than that. probably test set is different than train set?\n\nhttps://www.kaggle.com/code/miwojc/ultra-mnist-fastai-starter",
      "votes": null
    },
    {
      "id": "1730371",
      "postDate": "03/21/2022 07:30:56",
      "content": "<p>Are we allowed to use external pretrained models and external datasets?</p>",
      "rawMarkdown": "Are we allowed to use external pretrained models and external datasets?",
      "votes": null
    },
    {
      "id": "1730429",
      "postDate": "03/21/2022 08:38:25",
      "content": "<p>Yes, it is allowed. Please refer rules page. </p>\n<p>Thanks </p>",
      "rawMarkdown": "Yes, it is allowed. Please refer rules page. \n\nThanks",
      "votes": null
    },
    {
      "id": "1731113",
      "postDate": "03/22/2022 02:09:42",
      "content": "<p><a href=\"https://www.kaggle.com/miwojc\" target=\"_blank\">@miwojc</a> Isn't your validation score in notebook ~20% </p>",
      "rawMarkdown": "miwojc Isn't your validation score in notebook ~20%",
      "votes": null
    },
    {
      "id": "1731114",
      "postDate": "03/22/2022 02:10:00",
      "content": "<blockquote>\n  <p><a href=\"https://www.kaggle.com/miwojc\" target=\"_blank\">@miwojc</a> Isn't your validation score in notebook ~20%</p>\n</blockquote>\n<p>What is lb score for it</p>",
      "rawMarkdown": "> @miwojc Isn't your validation score in notebook ~20%\n\nWhat is lb score for it",
      "votes": null
    },
    {
      "id": "1733465",
      "postDate": "03/24/2022 09:55:49",
      "content": "<p>my initial submissions were half of the score on leaderboard as compared to validation score</p>",
      "rawMarkdown": "my initial submissions were half of the score on leaderboard as compared to validation score",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1730371,
      "author_name": "devanshchowdhury",
      "author_url": "",
      "post_date": "03/21/2022 07:30:56",
      "content": "<p>Are we allowed to use external pretrained models and external datasets?</p>",
      "votes": null,
      "replies": [
        {
          "id": 1730429,
          "author_name": "ubamba98",
          "author_url": "",
          "post_date": "03/21/2022 08:38:25",
          "content": "<p>Yes, it is allowed. Please refer rules page. </p>\n<p>Thanks </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1731113,
      "author_name": "l0new0lf",
      "author_url": "",
      "post_date": "03/22/2022 02:09:42",
      "content": "<p><a href=\"https://www.kaggle.com/miwojc\" target=\"_blank\">@miwojc</a> Isn't your validation score in notebook ~20% </p>",
      "votes": null,
      "replies": [
        {
          "id": 1731114,
          "author_name": "l0new0lf",
          "author_url": "",
          "post_date": "03/22/2022 02:10:00",
          "content": "<blockquote>\n  <p><a href=\"https://www.kaggle.com/miwojc\" target=\"_blank\">@miwojc</a> Isn't your validation score in notebook ~20%</p>\n</blockquote>\n<p>What is lb score for it</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1733465,
          "author_name": "miwojc",
          "author_url": "",
          "post_date": "03/24/2022 09:55:49",
          "content": "<p>my initial submissions were half of the score on leaderboard as compared to validation score</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1725466": "With resent and fastai local validation score is fine (at least for this stage in comp), however as hinted in this discussion (https://www.kaggle.com/c/ultra-mnist/discussion/313437), lb score is a lot lower than that. probably test set is different than train set?\n\nhttps://www.kaggle.com/code/miwojc/ultra-mnist-fastai-starter",
    "1730371": "Are we allowed to use external pretrained models and external datasets?",
    "1730429": "Yes, it is allowed. Please refer rules page. \n\nThanks",
    "1731113": "miwojc Isn't your validation score in notebook ~20%",
    "1731114": "> @miwojc Isn't your validation score in notebook ~20%\n\nWhat is lb score for it",
    "1733465": "my initial submissions were half of the score on leaderboard as compared to validation score"
  },
  "source": "meta"
}