{
  "id": 341194,
  "title": "Notebook Threw Exception",
  "url": "/competitions/hubmap-organ-segmentation/discussion/341194",
  "author_name": "",
  "post_date": "2022-08-01T18:13:03.401848500Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi!<br>\nCan anybody help me with debugging Notebook Threw Exception<br>\n<a href=\"https://www.kaggle.com/code/mcduck/baseline\" target=\"_blank\">https://www.kaggle.com/code/mcduck/baseline</a></p>\n<p>I tested it with train folder and it is ok, but submission fails</p>",
  "messages": [
    {
      "id": "1880447",
      "postDate": "08/01/2022 18:13:03",
      "content": "<p>Hi!<br>\nCan anybody help me with debugging Notebook Threw Exception<br>\n<a href=\"https://www.kaggle.com/code/mcduck/baseline\" target=\"_blank\">https://www.kaggle.com/code/mcduck/baseline</a></p>\n<p>I tested it with train folder and it is ok, but submission fails</p>",
      "rawMarkdown": "Hi!\nCan anybody help me with debugging Notebook Threw Exception\nhttps://www.kaggle.com/code/mcduck/baseline\n\nI tested it with train folder and it is ok, but submission fails",
      "votes": null
    },
    {
      "id": "1882370",
      "postDate": "08/03/2022 08:37:07",
      "content": "<p>Maybe it will be helpful for somebody, my mistake was on small images with shape less then (1024,1024). I create batch with only one image and squeeze batch dimension, so model eval fails</p>",
      "rawMarkdown": "Maybe it will be helpful for somebody, my mistake was on small images with shape less then (1024,1024). I create batch with only one image and squeeze batch dimension, so model eval fails",
      "votes": null
    },
    {
      "id": "1884073",
      "postDate": "08/04/2022 09:18:25",
      "content": "<p><a href=\"https://www.kaggle.com/mcduck\" target=\"_blank\">@mcduck</a> try reducing the batch size, I had the same issue where the inference notebook ran fine but threw exception after submitting. It was because the GPU memory was running out on the hidden test set.</p>",
      "rawMarkdown": "mcduck try reducing the batch size, I had the same issue where the inference notebook ran fine but threw exception after submitting. It was because the GPU memory was running out on the hidden test set.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1882370,
      "author_name": "mcduck",
      "author_url": "",
      "post_date": "08/03/2022 08:37:07",
      "content": "<p>Maybe it will be helpful for somebody, my mistake was on small images with shape less then (1024,1024). I create batch with only one image and squeeze batch dimension, so model eval fails</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1884073,
      "author_name": "yovinyahathugoda",
      "author_url": "",
      "post_date": "08/04/2022 09:18:25",
      "content": "<p><a href=\"https://www.kaggle.com/mcduck\" target=\"_blank\">@mcduck</a> try reducing the batch size, I had the same issue where the inference notebook ran fine but threw exception after submitting. It was because the GPU memory was running out on the hidden test set.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1880447": "Hi!\nCan anybody help me with debugging Notebook Threw Exception\nhttps://www.kaggle.com/code/mcduck/baseline\n\nI tested it with train folder and it is ok, but submission fails",
    "1882370": "Maybe it will be helpful for somebody, my mistake was on small images with shape less then (1024,1024). I create batch with only one image and squeeze batch dimension, so model eval fails",
    "1884073": "mcduck try reducing the batch size, I had the same issue where the inference notebook ran fine but threw exception after submitting. It was because the GPU memory was running out on the hidden test set."
  },
  "source": "meta"
}