{
  "id": 95753,
  "title": "Output Values for test data to continue experimenting",
  "url": "/competitions/freesound-audio-tagging-2019/discussion/95753",
  "author_name": "",
  "post_date": "2019-06-14T19:11:08.041364300Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi all, this was my first participation in Kaggle competition and had a great learning experience in last few weeks. I would like to continue working on this as an exercise offline to keep improving my model. Is there a way I can calculate the overall lwlrap score on test data? This way I can use my local LB score as a benchmark for my model.</p>",
  "messages": [
    {
      "id": "552897",
      "postDate": "06/14/2019 19:11:08",
      "content": "<p>Hi all, this was my first participation in Kaggle competition and had a great learning experience in last few weeks. I would like to continue working on this as an exercise offline to keep improving my model. Is there a way I can calculate the overall lwlrap score on test data? This way I can use my local LB score as a benchmark for my model.</p>",
      "rawMarkdown": "Hi all, this was my first participation in Kaggle competition and had a great learning experience in last few weeks. I would like to continue working on this as an exercise offline to keep improving my model. Is there a way I can calculate the overall lwlrap score on test data? This way I can use my local LB score as a benchmark for my model.",
      "votes": null
    },
    {
      "id": "553405",
      "postDate": "06/15/2019 16:23:49",
      "content": "<p>Hi Acku, I believe that after the final results are made public, you will be able to submit again to this competition, to continue experimenting.</p>\n\n<p>Also, in the future we will release the FSDKaggle2019 dataset used in this competition (including test set + ground truth), for the research community to use it beyond this competition.</p>",
      "rawMarkdown": "Hi Acku, I believe that after the final results are made public, you will be able to submit again to this competition, to continue experimenting.\n\nAlso, in the future we will release the FSDKaggle2019 dataset used in this competition (including test set + ground truth), for the research community to use it beyond this competition.",
      "votes": null
    },
    {
      "id": "554241",
      "postDate": "06/17/2019 07:20:26",
      "content": "<p>Hi Eduardo,\nThank you very much for the information.\nRegards</p>",
      "rawMarkdown": "Hi Eduardo,\nThank you very much for the information.\nRegards",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 553405,
      "author_name": "eduardofonseca",
      "author_url": "",
      "post_date": "06/15/2019 16:23:49",
      "content": "<p>Hi Acku, I believe that after the final results are made public, you will be able to submit again to this competition, to continue experimenting.</p>\n\n<p>Also, in the future we will release the FSDKaggle2019 dataset used in this competition (including test set + ground truth), for the research community to use it beyond this competition.</p>",
      "votes": null,
      "replies": [
        {
          "id": 554241,
          "author_name": "ackusingh",
          "author_url": "",
          "post_date": "06/17/2019 07:20:26",
          "content": "<p>Hi Eduardo,\nThank you very much for the information.\nRegards</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "552897": "Hi all, this was my first participation in Kaggle competition and had a great learning experience in last few weeks. I would like to continue working on this as an exercise offline to keep improving my model. Is there a way I can calculate the overall lwlrap score on test data? This way I can use my local LB score as a benchmark for my model.",
    "553405": "Hi Acku, I believe that after the final results are made public, you will be able to submit again to this competition, to continue experimenting.\n\nAlso, in the future we will release the FSDKaggle2019 dataset used in this competition (including test set + ground truth), for the research community to use it beyond this competition.",
    "554241": "Hi Eduardo,\nThank you very much for the information.\nRegards"
  },
  "source": "meta"
}