{
  "id": 565826,
  "title": "Validation/Test sequence sets are the same. Why?",
  "url": "/competitions/stanford-rna-3d-folding/discussion/565826",
  "author_name": "",
  "post_date": "2025-03-02T09:40:17.334503300Z",
  "votes": 3,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi, it seems like the val and test sequences.csv files are identical. Why sharing a toy val dataset if it is the same ? is it just to give us an example of a test set with and without its labels to test our pipelines ?<br>\nMaybe I am missing something, feel free to correct me if I am wrong :)</p>",
  "messages": [
    {
      "id": "3138150",
      "postDate": "03/02/2025 09:40:17",
      "content": "<p>Hi, it seems like the val and test sequences.csv files are identical. Why sharing a toy val dataset if it is the same ? is it just to give us an example of a test set with and without its labels to test our pipelines ?<br>\nMaybe I am missing something, feel free to correct me if I am wrong :)</p>",
      "rawMarkdown": "Hi, it seems like the val and test sequences.csv files are identical. Why sharing a toy val dataset if it is the same ? is it just to give us an example of a test set with and without its labels to test our pipelines ?\nMaybe I am missing something, feel free to correct me if I am wrong :)",
      "votes": null
    },
    {
      "id": "3148527",
      "postDate": "03/13/2025 08:40:51",
      "content": "<p>Yes, currently the validation and test sets are identical so you can fully test your pipeline (validation has labels, test does not). However, the test dataset will be updated in a future phase. After that update, the newly released test sequences (without labels) will be used for the official final evaluation, and the current public test set will move into training data. This setup ensures your pipeline can be verified now and then rerun on the updated test set later for final results.</p>",
      "rawMarkdown": "Yes, currently the validation and test sets are identical so you can fully test your pipeline (validation has labels, test does not). However, the test dataset will be updated in a future phase. After that update, the newly released test sequences (without labels) will be used for the official final evaluation, and the current public test set will move into training data. This setup ensures your pipeline can be verified now and then rerun on the updated test set later for final results.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3148527,
      "author_name": "younusmohamed",
      "author_url": "",
      "post_date": "03/13/2025 08:40:51",
      "content": "<p>Yes, currently the validation and test sets are identical so you can fully test your pipeline (validation has labels, test does not). However, the test dataset will be updated in a future phase. After that update, the newly released test sequences (without labels) will be used for the official final evaluation, and the current public test set will move into training data. This setup ensures your pipeline can be verified now and then rerun on the updated test set later for final results.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3138150": "Hi, it seems like the val and test sequences.csv files are identical. Why sharing a toy val dataset if it is the same ? is it just to give us an example of a test set with and without its labels to test our pipelines ?\nMaybe I am missing something, feel free to correct me if I am wrong :)",
    "3148527": "Yes, currently the validation and test sets are identical so you can fully test your pipeline (validation has labels, test does not). However, the test dataset will be updated in a future phase. After that update, the newly released test sequences (without labels) will be used for the official final evaluation, and the current public test set will move into training data. This setup ensures your pipeline can be verified now and then rerun on the updated test set later for final results."
  },
  "source": "meta"
}