{
  "id": 433412,
  "title": "Clarification on Final validation",
  "url": "/competitions/cafa-5-protein-function-prediction/discussion/433412",
  "author_name": "",
  "post_date": "2023-08-21T16:42:05.977743700Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>A bit late to ask this question, but I noticed that the evaluator uses only the top500 GO terms to validate a model. </p>\n<p>So, correct me if I'm wrong, but from my understanding, the current leaderboard is based on top500 on 50% of the test data, with the final results based on 100% of the test data. </p>\n<p>Will the final results still be based on top500 or are you planning to expand to all GO terms?</p>\n<p>Thanks!</p>",
  "messages": [
    {
      "id": "2401537",
      "postDate": "08/21/2023 16:42:05",
      "content": "<p>A bit late to ask this question, but I noticed that the evaluator uses only the top500 GO terms to validate a model. </p>\n<p>So, correct me if I'm wrong, but from my understanding, the current leaderboard is based on top500 on 50% of the test data, with the final results based on 100% of the test data. </p>\n<p>Will the final results still be based on top500 or are you planning to expand to all GO terms?</p>\n<p>Thanks!</p>",
      "rawMarkdown": "A bit late to ask this question, but I noticed that the evaluator uses only the top500 GO terms to validate a model. \n\nSo, correct me if I'm wrong, but from my understanding, the current leaderboard is based on top500 on 50% of the test data, with the final results based on 100% of the test data. \n\nWill the final results still be based on top500 or are you planning to expand to all GO terms?\n\nThanks!",
      "votes": null
    },
    {
      "id": "2401559",
      "postDate": "08/21/2023 16:52:53",
      "content": "<p>some of notebooks are based on top500 some of them not. For example my model based on top1500 go_id for each ontology. I belive that some people could make them model on local computer and then just upload results :) It's common to use top500 because with all go_id there is problem with memory and time :P </p>",
      "rawMarkdown": "some of notebooks are based on top500 some of them not. For example my model based on top1500 go_id for each ontology. I belive that some people could make them model on local computer and then just upload results :) It's common to use top500 because with all go_id there is problem with memory and time :P",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2401559,
      "author_name": "michalinahulak",
      "author_url": "",
      "post_date": "08/21/2023 16:52:53",
      "content": "<p>some of notebooks are based on top500 some of them not. For example my model based on top1500 go_id for each ontology. I belive that some people could make them model on local computer and then just upload results :) It's common to use top500 because with all go_id there is problem with memory and time :P </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2401537": "A bit late to ask this question, but I noticed that the evaluator uses only the top500 GO terms to validate a model. \n\nSo, correct me if I'm wrong, but from my understanding, the current leaderboard is based on top500 on 50% of the test data, with the final results based on 100% of the test data. \n\nWill the final results still be based on top500 or are you planning to expand to all GO terms?\n\nThanks!",
    "2401559": "some of notebooks are based on top500 some of them not. For example my model based on top1500 go_id for each ontology. I belive that some people could make them model on local computer and then just upload results :) It's common to use top500 because with all go_id there is problem with memory and time :P"
  },
  "source": "meta"
}