{
  "id": 207870,
  "title": "Model Accuracy Thread?",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/207870",
  "author_name": "",
  "post_date": "2020-12-31T17:31:04.422945600Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hey, newbie here!</p>\n<p>I was wondering if people could post their <code>validation accuracy</code> of their models as well as the <code>validation split</code> from train data, and their <em>LB-score</em> (if you have made a submission) to provide a way for everyone to quickly find a rough measure of how well their models are performing.</p>\n<p>This would be great to provide an idea of approximately how much rank you would get if you made a submission with equivalent metrics!</p>\n<p>Regards,<br>\nNeel Gupta</p>",
  "messages": [
    {
      "id": "1133956",
      "postDate": "12/31/2020 17:31:04",
      "content": "<p>Hey, newbie here!</p>\n<p>I was wondering if people could post their <code>validation accuracy</code> of their models as well as the <code>validation split</code> from train data, and their <em>LB-score</em> (if you have made a submission) to provide a way for everyone to quickly find a rough measure of how well their models are performing.</p>\n<p>This would be great to provide an idea of approximately how much rank you would get if you made a submission with equivalent metrics!</p>\n<p>Regards,<br>\nNeel Gupta</p>",
      "rawMarkdown": "Hey, newbie here!\n\nI was wondering if people could post their `validation accuracy` of their models as well as the `validation split` from train data, and their *LB-score* (if you have made a submission) to provide a way for everyone to quickly find a rough measure of how well their models are performing.\n\nThis would be great to provide an idea of approximately how much rank you would get if you made a submission with equivalent metrics!\n\nRegards,\nNeel Gupta",
      "votes": null
    },
    {
      "id": "1134238",
      "postDate": "01/01/2021 03:40:31",
      "content": "<p>This <a href=\"https://www.kaggle.com/c/cassava-leaf-disease-classification/discussion/203111\" target=\"_blank\">already exists</a></p>",
      "rawMarkdown": "This [already exists](https://www.kaggle.com/c/cassava-leaf-disease-classification/discussion/203111)",
      "votes": null
    },
    {
      "id": "1134496",
      "postDate": "01/01/2021 10:54:44",
      "content": "<p>yeah, ik but they do not post the accuracy score in <em>Percentage</em> rather they just tell the LB score. I was looking for someone to post their validation accuracy as well as the <code>split</code> so that atleast I have a rough idea of how much performance my model can give….</p>",
      "rawMarkdown": "yeah, ik but they do not post the accuracy score in *Percentage* rather they just tell the LB score. I was looking for someone to post their validation accuracy as well as the `split` so that atleast I have a rough idea of how much performance my model can give....",
      "votes": null
    },
    {
      "id": "1135083",
      "postDate": "01/01/2021 21:45:10",
      "content": "<p>If you want the accuracy score in percentage, just multiply by 100. For example the top team has a accuracy of 90.9% on the LB. People will often give their CV score along with their LB score. In the linked thread, for example,  someone mentioned their 5fold CV is 0.892 and their LB is 0.898. Hope this helps!</p>",
      "rawMarkdown": "If you want the accuracy score in percentage, just multiply by 100. For example the top team has a accuracy of 90.9% on the LB. People will often give their CV score along with their LB score. In the linked thread, for example,  someone mentioned their 5fold CV is 0.892 and their LB is 0.898. Hope this helps!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1134238,
      "author_name": "tanlikesmath",
      "author_url": "",
      "post_date": "01/01/2021 03:40:31",
      "content": "<p>This <a href=\"https://www.kaggle.com/c/cassava-leaf-disease-classification/discussion/203111\" target=\"_blank\">already exists</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 1134496,
          "author_name": "neelg007",
          "author_url": "",
          "post_date": "01/01/2021 10:54:44",
          "content": "<p>yeah, ik but they do not post the accuracy score in <em>Percentage</em> rather they just tell the LB score. I was looking for someone to post their validation accuracy as well as the <code>split</code> so that atleast I have a rough idea of how much performance my model can give….</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1135083,
          "author_name": "tanlikesmath",
          "author_url": "",
          "post_date": "01/01/2021 21:45:10",
          "content": "<p>If you want the accuracy score in percentage, just multiply by 100. For example the top team has a accuracy of 90.9% on the LB. People will often give their CV score along with their LB score. In the linked thread, for example,  someone mentioned their 5fold CV is 0.892 and their LB is 0.898. Hope this helps!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1133956": "Hey, newbie here!\n\nI was wondering if people could post their `validation accuracy` of their models as well as the `validation split` from train data, and their *LB-score* (if you have made a submission) to provide a way for everyone to quickly find a rough measure of how well their models are performing.\n\nThis would be great to provide an idea of approximately how much rank you would get if you made a submission with equivalent metrics!\n\nRegards,\nNeel Gupta",
    "1134238": "This [already exists](https://www.kaggle.com/c/cassava-leaf-disease-classification/discussion/203111)",
    "1134496": "yeah, ik but they do not post the accuracy score in *Percentage* rather they just tell the LB score. I was looking for someone to post their validation accuracy as well as the `split` so that atleast I have a rough idea of how much performance my model can give....",
    "1135083": "If you want the accuracy score in percentage, just multiply by 100. For example the top team has a accuracy of 90.9% on the LB. People will often give their CV score along with their LB score. In the linked thread, for example,  someone mentioned their 5fold CV is 0.892 and their LB is 0.898. Hope this helps!"
  },
  "source": "meta"
}