{
  "id": 520920,
  "title": "What's the right format for the output?",
  "url": "/competitions/rsna-2024-lumbar-spine-degenerative-classification/discussion/520920",
  "author_name": "samu2505",
  "post_date": "2024-07-18T01:49:17.286000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Maybe a dumb question, but what's the right format for the output? <br>\nDo we need to make predictions like this <code>25 * 3</code> as shown in this popular notebook <a href=\"https://www.kaggle.com/code/itsuki9180/rsna2024-lsdc-training-baseline\" target=\"_blank\">rsna2024-lsdc-training-baseline</a> or do we melt the dataframe with the 25 labels like this <code>df.melt(id_vars='study_id', var_name='condition_level', value_name='severity')</code> and then make predictions of the severity. <br>\nSo far, I am using the 2nd option, and I got 0.84 on the LB here is the notebook; <a href=\"https://www.kaggle.com/code/samu2505/rsna-pytorch-train-lb-0-84-cv-0-54\" target=\"_blank\">rsna-pytorch-train-lb-0-84-cv-0-54</a>.</p>",
  "messages": [
    {
      "id": 2926760,
      "postDate": "2024-07-18T01:49:17.287Z",
      "content": "<p>Maybe a dumb question, but what's the right format for the output? <br>\nDo we need to make predictions like this <code>25 * 3</code> as shown in this popular notebook <a href=\"https://www.kaggle.com/code/itsuki9180/rsna2024-lsdc-training-baseline\" target=\"_blank\">rsna2024-lsdc-training-baseline</a> or do we melt the dataframe with the 25 labels like this <code>df.melt(id_vars='study_id', var_name='condition_level', value_name='severity')</code> and then make predictions of the severity. <br>\nSo far, I am using the 2nd option, and I got 0.84 on the LB here is the notebook; <a href=\"https://www.kaggle.com/code/samu2505/rsna-pytorch-train-lb-0-84-cv-0-54\" target=\"_blank\">rsna-pytorch-train-lb-0-84-cv-0-54</a>.</p>",
      "rawMarkdown": "Maybe a dumb question, but what's the right format for the output? \nDo we need to make predictions like this `25 * 3` as shown in this popular notebook [rsna2024-lsdc-training-baseline](https://www.kaggle.com/code/itsuki9180/rsna2024-lsdc-training-baseline) or do we melt the dataframe with the 25 labels like this `df.melt(id_vars='study_id', var_name='condition_level', value_name='severity')` and then make predictions of the severity. \nSo far, I am using the 2nd option, and I got 0.84 on the LB here is the notebook; [rsna-pytorch-train-lb-0-84-cv-0-54](https://www.kaggle.com/code/samu2505/rsna-pytorch-train-lb-0-84-cv-0-54).",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2926760": "Maybe a dumb question, but what's the right format for the output? \nDo we need to make predictions like this `25 * 3` as shown in this popular notebook [rsna2024-lsdc-training-baseline](https://www.kaggle.com/code/itsuki9180/rsna2024-lsdc-training-baseline) or do we melt the dataframe with the 25 labels like this `df.melt(id_vars='study_id', var_name='condition_level', value_name='severity')` and then make predictions of the severity. \nSo far, I am using the 2nd option, and I got 0.84 on the LB here is the notebook; [rsna-pytorch-train-lb-0-84-cv-0-54](https://www.kaggle.com/code/samu2505/rsna-pytorch-train-lb-0-84-cv-0-54)."
  }
}