{
  "id": 159035,
  "title": "Is the high score ensemble kernel can running at private dataset?",
  "url": "/competitions/jigsaw-multilingual-toxic-comment-classification/discussion/159035",
  "author_name": "",
  "post_date": "2020-06-16T07:54:13.907268900Z",
  "votes": 3,
  "comment_count": 10,
  "views": 0,
  "content": "<p>Why 400+ folks in the ensemble kernel. How is the final score calculated in this game?\nI'm new to here, need help! please! </p>",
  "messages": [
    {
      "id": "888227",
      "postDate": "06/16/2020 07:54:13",
      "content": "<p>Why 400+ folks in the ensemble kernel. How is the final score calculated in this game?\nI'm new to here, need help! please! </p>",
      "rawMarkdown": "Why 400+ folks in the ensemble kernel. How is the final score calculated in this game?\nI'm new to here, need help! please!",
      "votes": null
    },
    {
      "id": "888265",
      "postDate": "06/16/2020 08:37:55",
      "content": "<p>yes,even though it is a code competition but still for this competition still you can just upload csv to kernel and submit from kernel.</p>\n\n<p>Confirmation: <a href=\"https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/141554#800703\">https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/141554#800703</a></p>\n\n<p>it will work for private lb too as we don't  have  any hidden private test set in this competition</p>\n\n<p>check this post and it's comments as well : <a href=\"https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/147331\">https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/147331</a></p>",
      "rawMarkdown": "yes,even though it is a code competition but still for this competition still you can just upload csv to kernel and submit from kernel.\n\nConfirmation: https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/141554#800703\n\n\nit will work for private lb too as we don't  have  any hidden private test set in this competition\n\ncheck this post and it's comments as well : https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/147331",
      "votes": null
    },
    {
      "id": "888284",
      "postDate": "06/16/2020 08:58:48",
      "content": "<p>So, the test dataset already contains the private dataset and only evaluate part of it at the stage 1, right?</p>",
      "rawMarkdown": "So, the test dataset already contains the private dataset and only evaluate part of it at the stage 1, right?",
      "votes": null
    },
    {
      "id": "888290",
      "postDate": "06/16/2020 09:01:13",
      "content": "<p>yes, correct <a href=\"/majunhua\">@majunhua</a> </p>",
      "rawMarkdown": "yes, correct @majunhua",
      "votes": null
    },
    {
      "id": "888296",
      "postDate": "06/16/2020 09:03:57",
      "content": "<p>Thank you for your patience，Mobassir。</p>",
      "rawMarkdown": "Thank you for your patience，Mobassir。",
      "votes": null
    },
    {
      "id": "888404",
      "postDate": "06/16/2020 10:17:18",
      "content": "<p>Since the whole test dataset is available (both public and private parts), whatever you predict on it offline will be the same thing online thus you can use the <code>submission.csv</code> file that you create offline. </p>\n\n<p>That's not always the case in other competition where only a fraction of the test dataset is provided and thus the <code>submission.csv</code> should be generated programmatically (i.e. from your submitted notebook). </p>\n\n<p>I hope this makes things clearer.  </p>",
      "rawMarkdown": "Since the whole test dataset is available (both public and private parts), whatever you predict on it offline will be the same thing online thus you can use the `submission.csv` file that you create offline. \n\nThat's not always the case in other competition where only a fraction of the test dataset is provided and thus the `submission.csv` should be generated programmatically (i.e. from your submitted notebook). \n\nI hope this makes things clearer.",
      "votes": null
    },
    {
      "id": "889455",
      "postDate": "06/17/2020 00:59:22",
      "content": "<p><a href=\"/mobassir\">@mobassir</a> Does this mean that one can directly use offline generated submision.csv for private lb too? I mean, is it not required to have a model generate the required csv within the kernel for private ranking?</p>",
      "rawMarkdown": "mobassir Does this mean that one can directly use offline generated submision.csv for private lb too? I mean, is it not required to have a model generate the required csv within the kernel for private ranking?",
      "votes": null
    },
    {
      "id": "891218",
      "postDate": "06/18/2020 02:56:51",
      "content": "<p>You can do the prediction offline, save in a csv, upload it into kaggle and submit it. (it work only with this competition)</p>",
      "rawMarkdown": "You can do the prediction offline, save in a csv, upload it into kaggle and submit it. (it work only with this competition)",
      "votes": null
    },
    {
      "id": "891348",
      "postDate": "06/18/2020 05:41:55",
      "content": "<p><a href=\"/ankitsajwan\">@ankitsajwan</a>  sorry  i didn't notice your reply,,yes <a href=\"/ludovick\">@ludovick</a>   is right</p>",
      "rawMarkdown": "ankitsajwan  sorry  i didn't notice your reply,,yes @ludovick   is right",
      "votes": null
    },
    {
      "id": "891659",
      "postDate": "06/18/2020 11:09:48",
      "content": "<p>Thank you, Yassine, it help me a lot.</p>",
      "rawMarkdown": "Thank you, Yassine, it help me a lot.",
      "votes": null
    },
    {
      "id": "891662",
      "postDate": "06/18/2020 11:11:57",
      "content": "<p>Yes, you are right！</p>",
      "rawMarkdown": "Yes, you are right！",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 888265,
      "author_name": "mobassir",
      "author_url": "",
      "post_date": "06/16/2020 08:37:55",
      "content": "<p>yes,even though it is a code competition but still for this competition still you can just upload csv to kernel and submit from kernel.</p>\n\n<p>Confirmation: <a href=\"https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/141554#800703\">https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/141554#800703</a></p>\n\n<p>it will work for private lb too as we don't  have  any hidden private test set in this competition</p>\n\n<p>check this post and it's comments as well : <a href=\"https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/147331\">https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/147331</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 888284,
          "author_name": "majunhua",
          "author_url": "",
          "post_date": "06/16/2020 08:58:48",
          "content": "<p>So, the test dataset already contains the private dataset and only evaluate part of it at the stage 1, right?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 888290,
          "author_name": "mobassir",
          "author_url": "",
          "post_date": "06/16/2020 09:01:13",
          "content": "<p>yes, correct <a href=\"/majunhua\">@majunhua</a> </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 888296,
          "author_name": "majunhua",
          "author_url": "",
          "post_date": "06/16/2020 09:03:57",
          "content": "<p>Thank you for your patience，Mobassir。</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 889455,
          "author_name": "ankitsajwan",
          "author_url": "",
          "post_date": "06/17/2020 00:59:22",
          "content": "<p><a href=\"/mobassir\">@mobassir</a> Does this mean that one can directly use offline generated submision.csv for private lb too? I mean, is it not required to have a model generate the required csv within the kernel for private ranking?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 891218,
          "author_name": "ludovick",
          "author_url": "",
          "post_date": "06/18/2020 02:56:51",
          "content": "<p>You can do the prediction offline, save in a csv, upload it into kaggle and submit it. (it work only with this competition)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 891348,
          "author_name": "mobassir",
          "author_url": "",
          "post_date": "06/18/2020 05:41:55",
          "content": "<p><a href=\"/ankitsajwan\">@ankitsajwan</a>  sorry  i didn't notice your reply,,yes <a href=\"/ludovick\">@ludovick</a>   is right</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 891662,
          "author_name": "majunhua",
          "author_url": "",
          "post_date": "06/18/2020 11:11:57",
          "content": "<p>Yes, you are right！</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 888404,
      "author_name": "yassinealouini",
      "author_url": "",
      "post_date": "06/16/2020 10:17:18",
      "content": "<p>Since the whole test dataset is available (both public and private parts), whatever you predict on it offline will be the same thing online thus you can use the <code>submission.csv</code> file that you create offline. </p>\n\n<p>That's not always the case in other competition where only a fraction of the test dataset is provided and thus the <code>submission.csv</code> should be generated programmatically (i.e. from your submitted notebook). </p>\n\n<p>I hope this makes things clearer.  </p>",
      "votes": null,
      "replies": [
        {
          "id": 891659,
          "author_name": "majunhua",
          "author_url": "",
          "post_date": "06/18/2020 11:09:48",
          "content": "<p>Thank you, Yassine, it help me a lot.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "888227": "Why 400+ folks in the ensemble kernel. How is the final score calculated in this game?\nI'm new to here, need help! please!",
    "888265": "yes,even though it is a code competition but still for this competition still you can just upload csv to kernel and submit from kernel.\n\nConfirmation: https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/141554#800703\n\n\nit will work for private lb too as we don't  have  any hidden private test set in this competition\n\ncheck this post and it's comments as well : https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/147331",
    "888284": "So, the test dataset already contains the private dataset and only evaluate part of it at the stage 1, right?",
    "888290": "yes, correct @majunhua",
    "888296": "Thank you for your patience，Mobassir。",
    "888404": "Since the whole test dataset is available (both public and private parts), whatever you predict on it offline will be the same thing online thus you can use the `submission.csv` file that you create offline. \n\nThat's not always the case in other competition where only a fraction of the test dataset is provided and thus the `submission.csv` should be generated programmatically (i.e. from your submitted notebook). \n\nI hope this makes things clearer.",
    "889455": "mobassir Does this mean that one can directly use offline generated submision.csv for private lb too? I mean, is it not required to have a model generate the required csv within the kernel for private ranking?",
    "891218": "You can do the prediction offline, save in a csv, upload it into kaggle and submit it. (it work only with this competition)",
    "891348": "ankitsajwan  sorry  i didn't notice your reply,,yes @ludovick   is right",
    "891659": "Thank you, Yassine, it help me a lot.",
    "891662": "Yes, you are right！"
  },
  "source": "meta"
}