{
  "id": 614021,
  "title": "Are Track 1 and Track 2 using the same private test set?",
  "url": "/competitions/2026-icassp-hyper-object-challenge-track-2/discussion/614021",
  "author_name": "",
  "post_date": "2025-10-31T12:38:44.636833Z",
  "votes": null,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Dear organizers,</p>\n<p>I noticed that in the 2026 ICASSP Hyper-Object Challenge, both Track 1 and Track 2 seem to use the same dataset for submission prediction.</p>\n<p>Specifically, the prediction files for both tracks have the same number of samples and identical data structure. This makes me wonder whether both tracks are sharing the same private test set.</p>\n<p>Could you please confirm: Are Track 1 and Track 2 indeed using the same private test data?</p>\n<p>I would like to make sure our experimental setup aligns correctly with the intended design of the challenge.</p>\n<p>Thank you very much for organizing this great competition!</p>",
  "messages": [
    {
      "id": "3309316",
      "postDate": "10/31/2025 12:38:44",
      "content": "<p>Dear organizers,</p>\n<p>I noticed that in the 2026 ICASSP Hyper-Object Challenge, both Track 1 and Track 2 seem to use the same dataset for submission prediction.</p>\n<p>Specifically, the prediction files for both tracks have the same number of samples and identical data structure. This makes me wonder whether both tracks are sharing the same private test set.</p>\n<p>Could you please confirm: Are Track 1 and Track 2 indeed using the same private test data?</p>\n<p>I would like to make sure our experimental setup aligns correctly with the intended design of the challenge.</p>\n<p>Thank you very much for organizing this great competition!</p>",
      "rawMarkdown": "Dear organizers,\n\nI noticed that in the 2026 ICASSP Hyper-Object Challenge, both Track 1 and Track 2 seem to use the same dataset for submission prediction.\n\nSpecifically, the prediction files for both tracks have the same number of samples and identical data structure. This makes me wonder whether both tracks are sharing the same private test set.\n\nCould you please confirm: Are Track 1 and Track 2 indeed using the same private test data?\n\nI would like to make sure our experimental setup aligns correctly with the intended design of the challenge.\n\nThank you very much for organizing this great competition!",
      "votes": null
    },
    {
      "id": "3313070",
      "postDate": "11/08/2025 15:42:51",
      "content": "<p>The private test data for Track 1 and Track 2 should be the same.</p>",
      "rawMarkdown": "The private test data for Track 1 and Track 2 should be the same.",
      "votes": null
    },
    {
      "id": "3313083",
      "postDate": "11/08/2025 16:09:23",
      "content": "<p>May I ask if  final score is calculated based on the current ranking table, or does need to upload the code and weights for the organizers to test?</p>",
      "rawMarkdown": "May I ask if  final score is calculated based on the current ranking table, or does need to upload the code and weights for the organizers to test?",
      "votes": null
    },
    {
      "id": "3313096",
      "postDate": "11/08/2025 16:38:13",
      "content": "<p>The final ranking will be verified based on the submitted code and weights.</p>",
      "rawMarkdown": "The final ranking will be verified based on the submitted code and weights.",
      "votes": null
    },
    {
      "id": "3313097",
      "postDate": "11/08/2025 16:43:07",
      "content": "<p>We are also a bit confused about the team limits. The rules mention a\nlimit of 5 members per team, but on Kaggle it seems possible to invite up to 9 members.What is the actual team member limit?Also, do professors or supervisors count toward the team member limit?</p>",
      "rawMarkdown": "We are also a bit confused about the team limits. The rules mention a\nlimit of 5 members per team, but on Kaggle it seems possible to invite up to 9 members.What is the actual team member limit?Also, do professors or supervisors count toward the team member limit?",
      "votes": null
    },
    {
      "id": "3313116",
      "postDate": "11/08/2025 17:28:09",
      "content": "<p>Kaggle platform is used as the playground for teams to work on the problem and get a sense of how the evaluation will be conducted. The rules in Kaggle are default rule which we didn't modify. On our end, there's not limit on the team size. We will update the rule. </p>\n<p>The final evaluation will be based on the hidden-private test set that we will share with registered team. The final evaluation will be verified on the submitted code and model weight later. Instructions will be provided on how to submit the code and trained model to the registered team.</p>",
      "rawMarkdown": "Kaggle platform is used as the playground for teams to work on the problem and get a sense of how the evaluation will be conducted. The rules in Kaggle are default rule which we didn't modify. On our end, there's not limit on the team size. We will update the rule. \n\nThe final evaluation will be based on the hidden-private test set that we will share with registered team. The final evaluation will be verified on the submitted code and model weight later. Instructions will be provided on how to submit the code and trained model to the registered team.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3313070,
      "author_name": "hyperobject",
      "author_url": "",
      "post_date": "11/08/2025 15:42:51",
      "content": "<p>The private test data for Track 1 and Track 2 should be the same.</p>",
      "votes": null,
      "replies": [
        {
          "id": 3313083,
          "author_name": "sunxioajun920",
          "author_url": "",
          "post_date": "11/08/2025 16:09:23",
          "content": "<p>May I ask if  final score is calculated based on the current ranking table, or does need to upload the code and weights for the organizers to test?</p>",
          "votes": null,
          "replies": [
            {
              "id": 3313096,
              "author_name": "hyperobject",
              "author_url": "",
              "post_date": "11/08/2025 16:38:13",
              "content": "<p>The final ranking will be verified based on the submitted code and weights.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 3313097,
                  "author_name": "sunxioajun920",
                  "author_url": "",
                  "post_date": "11/08/2025 16:43:07",
                  "content": "<p>We are also a bit confused about the team limits. The rules mention a\nlimit of 5 members per team, but on Kaggle it seems possible to invite up to 9 members.What is the actual team member limit?Also, do professors or supervisors count toward the team member limit?</p>",
                  "votes": null,
                  "replies": [
                    {
                      "id": 3313116,
                      "author_name": "hyperobject",
                      "author_url": "",
                      "post_date": "11/08/2025 17:28:09",
                      "content": "<p>Kaggle platform is used as the playground for teams to work on the problem and get a sense of how the evaluation will be conducted. The rules in Kaggle are default rule which we didn't modify. On our end, there's not limit on the team size. We will update the rule. </p>\n<p>The final evaluation will be based on the hidden-private test set that we will share with registered team. The final evaluation will be verified on the submitted code and model weight later. Instructions will be provided on how to submit the code and trained model to the registered team.</p>",
                      "votes": null,
                      "replies": []
                    }
                  ]
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3309316": "Dear organizers,\n\nI noticed that in the 2026 ICASSP Hyper-Object Challenge, both Track 1 and Track 2 seem to use the same dataset for submission prediction.\n\nSpecifically, the prediction files for both tracks have the same number of samples and identical data structure. This makes me wonder whether both tracks are sharing the same private test set.\n\nCould you please confirm: Are Track 1 and Track 2 indeed using the same private test data?\n\nI would like to make sure our experimental setup aligns correctly with the intended design of the challenge.\n\nThank you very much for organizing this great competition!",
    "3313070": "The private test data for Track 1 and Track 2 should be the same.",
    "3313083": "May I ask if  final score is calculated based on the current ranking table, or does need to upload the code and weights for the organizers to test?",
    "3313096": "The final ranking will be verified based on the submitted code and weights.",
    "3313097": "We are also a bit confused about the team limits. The rules mention a\nlimit of 5 members per team, but on Kaggle it seems possible to invite up to 9 members.What is the actual team member limit?Also, do professors or supervisors count toward the team member limit?",
    "3313116": "Kaggle platform is used as the playground for teams to work on the problem and get a sense of how the evaluation will be conducted. The rules in Kaggle are default rule which we didn't modify. On our end, there's not limit on the team size. We will update the rule. \n\nThe final evaluation will be based on the hidden-private test set that we will share with registered team. The final evaluation will be verified on the submitted code and model weight later. Instructions will be provided on how to submit the code and trained model to the registered team."
  },
  "source": "meta"
}