{
  "id": 88721,
  "title": "Found score without summission fast.ai",
  "url": "/competitions/freesound-audio-tagging-2019/discussion/88721",
  "author_name": "",
  "post_date": "2019-04-10T07:28:30.741822800Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Good morning, this is my first competition i would know if there is another way to find out the score of my kernel without a submission or if the only whay is study the learning range of the execution. </p>\n\n<p>thanks for your advice.</p>",
  "messages": [
    {
      "id": "511649",
      "postDate": "04/10/2019 07:28:30",
      "content": "<p>Good morning, this is my first competition i would know if there is another way to find out the score of my kernel without a submission or if the only whay is study the learning range of the execution. </p>\n\n<p>thanks for your advice.</p>",
      "rawMarkdown": "Good morning, this is my first competition i would know if there is another way to find out the score of my kernel without a submission or if the only whay is study the learning range of the execution. \n\nthanks for your advice.",
      "votes": null
    },
    {
      "id": "512225",
      "postDate": "04/10/2019 16:31:56",
      "content": "<p>You have to make a submission to know your evaluation score on the private test set.</p>\n\n<p>However, you can also set aside part of the curated training data as a held-out validation set and compute the evaluation metric yourself on this validation set. The evaluation page <a href=\"https://www.kaggle.com/c/freesound-audio-tagging-2019/overview/evaluation\">https://www.kaggle.com/c/freesound-audio-tagging-2019/overview/evaluation</a> has more details about the metric and links to a Colab notebook <a href=\"https://colab.research.google.com/drive/1AgPdhSp7ttY18O3fEoHOQKlt_3HJDLi8\">https://colab.research.google.com/drive/1AgPdhSp7ttY18O3fEoHOQKlt_3HJDLi8</a> with code that computes the metric. Note that this also allows you to compute per-class metrics so you can dig into which classes are performing better than others, in case that helps you to refine your model.</p>",
      "rawMarkdown": "You have to make a submission to know your evaluation score on the private test set.\n\nHowever, you can also set aside part of the curated training data as a held-out validation set and compute the evaluation metric yourself on this validation set. The evaluation page https://www.kaggle.com/c/freesound-audio-tagging-2019/overview/evaluation has more details about the metric and links to a Colab notebook https://colab.research.google.com/drive/1AgPdhSp7ttY18O3fEoHOQKlt_3HJDLi8 with code that computes the metric. Note that this also allows you to compute per-class metrics so you can dig into which classes are performing better than others, in case that helps you to refine your model.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 512225,
      "author_name": "plakal",
      "author_url": "",
      "post_date": "04/10/2019 16:31:56",
      "content": "<p>You have to make a submission to know your evaluation score on the private test set.</p>\n\n<p>However, you can also set aside part of the curated training data as a held-out validation set and compute the evaluation metric yourself on this validation set. The evaluation page <a href=\"https://www.kaggle.com/c/freesound-audio-tagging-2019/overview/evaluation\">https://www.kaggle.com/c/freesound-audio-tagging-2019/overview/evaluation</a> has more details about the metric and links to a Colab notebook <a href=\"https://colab.research.google.com/drive/1AgPdhSp7ttY18O3fEoHOQKlt_3HJDLi8\">https://colab.research.google.com/drive/1AgPdhSp7ttY18O3fEoHOQKlt_3HJDLi8</a> with code that computes the metric. Note that this also allows you to compute per-class metrics so you can dig into which classes are performing better than others, in case that helps you to refine your model.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "511649": "Good morning, this is my first competition i would know if there is another way to find out the score of my kernel without a submission or if the only whay is study the learning range of the execution. \n\nthanks for your advice.",
    "512225": "You have to make a submission to know your evaluation score on the private test set.\n\nHowever, you can also set aside part of the curated training data as a held-out validation set and compute the evaluation metric yourself on this validation set. The evaluation page https://www.kaggle.com/c/freesound-audio-tagging-2019/overview/evaluation has more details about the metric and links to a Colab notebook https://colab.research.google.com/drive/1AgPdhSp7ttY18O3fEoHOQKlt_3HJDLi8 with code that computes the metric. Note that this also allows you to compute per-class metrics so you can dig into which classes are performing better than others, in case that helps you to refine your model."
  },
  "source": "meta"
}