{
  "id": 203166,
  "title": "Averaging submissions ",
  "url": "/competitions/rfcx-species-audio-detection/discussion/203166",
  "author_name": "",
  "post_date": "2020-12-14T04:00:56.987338200Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Is it possible to averaging submissions from different Kernels or it makes error at the end of this competition </p>",
  "messages": [
    {
      "id": "1111806",
      "postDate": "12/14/2020 04:00:56",
      "content": "<p>Is it possible to averaging submissions from different Kernels or it makes error at the end of this competition </p>",
      "rawMarkdown": "Is it possible to averaging submissions from different Kernels or it makes error at the end of this competition",
      "votes": null
    },
    {
      "id": "1112441",
      "postDate": "12/14/2020 15:48:56",
      "content": "<p>You can ensamble the submissions as you want, I think!</p>\n<p>Best, </p>\n<p>Guglielmo</p>",
      "rawMarkdown": "You can ensamble the submissions as you want, I think!\n\nBest, \n\nGuglielmo",
      "votes": null
    },
    {
      "id": "1115355",
      "postDate": "12/16/2020 08:29:27",
      "content": "<p>You can see <a href=\"https://www.kaggle.com/duythanhng/all-in-one-rfcx-baseline-model-avg\" target=\"_blank\">this notebook</a> </p>",
      "rawMarkdown": "You can see [this notebook](https://www.kaggle.com/duythanhng/all-in-one-rfcx-baseline-model-avg)",
      "votes": null
    },
    {
      "id": "1119003",
      "postDate": "12/19/2020 16:13:59",
      "content": "<p>Yes , no problem in this comeptition , because you are just uploading .csv dataset so they are not getting access to your model just evaluating a score on csv file ,but sometimes on other comeptition you have to give your notebook so they actually run you code on hidden dataset which you dont know about , generate another submission , then if you are just averaging on your notebook  then it wont work as it wont get acess to the model to predict new data , if you are also creating the model then its just okay </p>",
      "rawMarkdown": "Yes , no problem in this comeptition , because you are just uploading .csv dataset so they are not getting access to your model just evaluating a score on csv file ,but sometimes on other comeptition you have to give your notebook so they actually run you code on hidden dataset which you dont know about , generate another submission , then if you are just averaging on your notebook  then it wont work as it wont get acess to the model to predict new data , if you are also creating the model then its just okay",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1112441,
      "author_name": "guglielmocamporese",
      "author_url": "",
      "post_date": "12/14/2020 15:48:56",
      "content": "<p>You can ensamble the submissions as you want, I think!</p>\n<p>Best, </p>\n<p>Guglielmo</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1115355,
      "author_name": "duythanhng",
      "author_url": "",
      "post_date": "12/16/2020 08:29:27",
      "content": "<p>You can see <a href=\"https://www.kaggle.com/duythanhng/all-in-one-rfcx-baseline-model-avg\" target=\"_blank\">this notebook</a> </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1119003,
      "author_name": "swaralipibose",
      "author_url": "",
      "post_date": "12/19/2020 16:13:59",
      "content": "<p>Yes , no problem in this comeptition , because you are just uploading .csv dataset so they are not getting access to your model just evaluating a score on csv file ,but sometimes on other comeptition you have to give your notebook so they actually run you code on hidden dataset which you dont know about , generate another submission , then if you are just averaging on your notebook  then it wont work as it wont get acess to the model to predict new data , if you are also creating the model then its just okay </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1111806": "Is it possible to averaging submissions from different Kernels or it makes error at the end of this competition",
    "1112441": "You can ensamble the submissions as you want, I think!\n\nBest, \n\nGuglielmo",
    "1115355": "You can see [this notebook](https://www.kaggle.com/duythanhng/all-in-one-rfcx-baseline-model-avg)",
    "1119003": "Yes , no problem in this comeptition , because you are just uploading .csv dataset so they are not getting access to your model just evaluating a score on csv file ,but sometimes on other comeptition you have to give your notebook so they actually run you code on hidden dataset which you dont know about , generate another submission , then if you are just averaging on your notebook  then it wont work as it wont get acess to the model to predict new data , if you are also creating the model then its just okay"
  },
  "source": "meta"
}