{
  "id": 621647,
  "title": "Question About the Accuracy of the Validation Labels",
  "url": "/competitions/the-3lc-cotton-weed-detection-challenge/discussion/621647",
  "author_name": "GaGaWuLaLa",
  "post_date": "2025-11-15T09:57:37.678000",
  "votes": 0,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi, I would like to clarify something about the dataset.\nAre the annotations in the validation set guaranteed to be correct, or should we expect that some validation labels may also contain errors just like the training set?</p>\n<p>Thanks!</p>",
  "messages": [
    {
      "id": 3332652,
      "postDate": "2025-11-16T19:46:13.890Z",
      "content": "<p>The annotations in the validation set were labeled/oriented manually by me. I can guarantee that majority of the labels are correctly labeled. There are a very few set of examples in the set which are ambiguous, overlapping, or partially visible where the labelling was done based on my instinct and what I thought was the right label. If you disagree with the assigned label please feel free to update those labels to whatever you believe is most accurate using the 3lc dashboard. Ping here if you face any issues while doing so.</p>",
      "rawMarkdown": "The annotations in the validation set were labeled/oriented manually by me. I can guarantee that majority of the labels are correctly labeled. There are a very few set of examples in the set which are ambiguous, overlapping, or partially visible where the labelling was done based on my instinct and what I thought was the right label. If you disagree with the assigned label please feel free to update those labels to whatever you believe is most accurate using the 3lc dashboard. Ping here if you face any issues while doing so."
    },
    {
      "id": 3331469,
      "postDate": "2025-11-16T13:07:19.297Z",
      "content": "<p>Hello Gagawulala, I’m expecting that the train and validation have some missing or incorrect labels. We must use 3LC tools to identify this and retrain. Another question is: the dataset has many similar plants like Malva and Ipomea, so I hope we need to identify these similar cases.</p>",
      "rawMarkdown": "Hello Gagawulala, I’m expecting that the train and validation have some missing or incorrect labels. We must use 3LC tools to identify this and retrain. Another question is: the dataset has many similar plants like Malva and Ipomea, so I hope we need to identify these similar cases."
    },
    {
      "id": 3326358,
      "postDate": "2025-11-15T09:57:37.680Z",
      "content": "<p>Hi, I would like to clarify something about the dataset.\nAre the annotations in the validation set guaranteed to be correct, or should we expect that some validation labels may also contain errors just like the training set?</p>\n<p>Thanks!</p>",
      "rawMarkdown": "Hi, I would like to clarify something about the dataset.\nAre the annotations in the validation set guaranteed to be correct, or should we expect that some validation labels may also contain errors just like the training set?\n\nThanks!"
    },
    {
      "id": 3331534,
      "postDate": "2025-11-16T13:21:34.550Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 3332652,
      "author_name": "",
      "author_url": "",
      "post_date": "2025-11-16T19:46:13.890000",
      "content": "<p>The annotations in the validation set were labeled/oriented manually by me. I can guarantee that majority of the labels are correctly labeled. There are a very few set of examples in the set which are ambiguous, overlapping, or partially visible where the labelling was done based on my instinct and what I thought was the right label. If you disagree with the assigned label please feel free to update those labels to whatever you believe is most accurate using the 3lc dashboard. Ping here if you face any issues while doing so.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 3331469,
      "author_name": "Gustavo Vicente",
      "author_url": "",
      "post_date": "2025-11-16T13:07:19.297000",
      "content": "<p>Hello Gagawulala, I’m expecting that the train and validation have some missing or incorrect labels. We must use 3LC tools to identify this and retrain. Another question is: the dataset has many similar plants like Malva and Ipomea, so I hope we need to identify these similar cases.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 3331534,
      "author_name": "",
      "author_url": "",
      "post_date": "2025-11-16T13:21:34.550000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3332652": "The annotations in the validation set were labeled/oriented manually by me. I can guarantee that majority of the labels are correctly labeled. There are a very few set of examples in the set which are ambiguous, overlapping, or partially visible where the labelling was done based on my instinct and what I thought was the right label. If you disagree with the assigned label please feel free to update those labels to whatever you believe is most accurate using the 3lc dashboard. Ping here if you face any issues while doing so.",
    "3331469": "Hello Gagawulala, I’m expecting that the train and validation have some missing or incorrect labels. We must use 3LC tools to identify this and retrain. Another question is: the dataset has many similar plants like Malva and Ipomea, so I hope we need to identify these similar cases.",
    "3326358": "Hi, I would like to clarify something about the dataset.\nAre the annotations in the validation set guaranteed to be correct, or should we expect that some validation labels may also contain errors just like the training set?\n\nThanks!",
    "3331534": ""
  }
}