{
  "id": 465764,
  "title": "Unable to submit predictions",
  "url": "/competitions/blood-vessel-segmentation/discussion/465764",
  "author_name": "",
  "post_date": "2024-01-05T14:02:21.812720Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi, everyone<br>\nOn my notebook, I am save both the model (in h5 format) and the submission.csv, but neither of those are recognized as output files by Kaggle. Hence, I can not make a submission for the competition. How did you submit your notebook? Note that I have one notebook for all tasks (data loading and preprocessing, model development and training and submitting predictions). Is it commendable to have one notebook for training and one for inference?</p>",
  "messages": [
    {
      "id": "2588446",
      "postDate": "01/05/2024 14:02:21",
      "content": "<p>Hi, everyone<br>\nOn my notebook, I am save both the model (in h5 format) and the submission.csv, but neither of those are recognized as output files by Kaggle. Hence, I can not make a submission for the competition. How did you submit your notebook? Note that I have one notebook for all tasks (data loading and preprocessing, model development and training and submitting predictions). Is it commendable to have one notebook for training and one for inference?</p>",
      "rawMarkdown": "Hi, everyone\nOn my notebook, I am save both the model (in h5 format) and the submission.csv, but neither of those are recognized as output files by Kaggle. Hence, I can not make a submission for the competition. How did you submit your notebook? Note that I have one notebook for all tasks (data loading and preprocessing, model development and training and submitting predictions). Is it commendable to have one notebook for training and one for inference?",
      "votes": null
    },
    {
      "id": "2588728",
      "postDate": "01/05/2024 17:21:20",
      "content": "<p>\"Is it commendable to have one notebook for training and one for inference?\"<br>\nYes, you have to submit an inference notebook that make a submission.csv file with the rle predictions of the files of a general test folder. I t will be executed on another environement.</p>",
      "rawMarkdown": "\"Is it commendable to have one notebook for training and one for inference?\"\nYes, you have to submit an inference notebook that make a submission.csv file with the rle predictions of the files of a general test folder. I t will be executed on another environement.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2588728,
      "author_name": "sacuscreed",
      "author_url": "",
      "post_date": "01/05/2024 17:21:20",
      "content": "<p>\"Is it commendable to have one notebook for training and one for inference?\"<br>\nYes, you have to submit an inference notebook that make a submission.csv file with the rle predictions of the files of a general test folder. I t will be executed on another environement.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2588446": "Hi, everyone\nOn my notebook, I am save both the model (in h5 format) and the submission.csv, but neither of those are recognized as output files by Kaggle. Hence, I can not make a submission for the competition. How did you submit your notebook? Note that I have one notebook for all tasks (data loading and preprocessing, model development and training and submitting predictions). Is it commendable to have one notebook for training and one for inference?",
    "2588728": "\"Is it commendable to have one notebook for training and one for inference?\"\nYes, you have to submit an inference notebook that make a submission.csv file with the rle predictions of the files of a general test folder. I t will be executed on another environement."
  },
  "source": "meta"
}