{
  "id": 371187,
  "title": "why forked public inference notebooks time out?",
  "url": "/competitions/rsna-breast-cancer-detection/discussion/371187",
  "author_name": "dragon zhang",
  "post_date": "2022-12-08T13:27:40.300000",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Is the kernel assigned with different hardware resources?</p>\n<p>Since the public shared notebooks with LB scores, why my forked learning ones time out?</p>",
  "messages": [
    {
      "id": 2059311,
      "postDate": "2022-12-08T17:41:47.833Z",
      "content": "<p>Which one is timing out?  Agreed, that is problematic for sure.  And yes, different hardware resources is a common issue which has been raised before.</p>\n<p>You could try using the faster dicom decode library, dicomsdl </p>\n<p>Note if Kaggle supplied us with lz4 compressed files of the pixel data instead of dicom, we'd see up to a ~10x speedup in decoding, though it would take about 2x the amount of space.</p>",
      "rawMarkdown": "Which one is timing out?  Agreed, that is problematic for sure.  And yes, different hardware resources is a common issue which has been raised before.\n\nYou could try using the faster dicom decode library, dicomsdl \n\nNote if Kaggle supplied us with lz4 compressed files of the pixel data instead of dicom, we'd see up to a ~10x speedup in decoding, though it would take about 2x the amount of space.",
      "votes": 1
    },
    {
      "id": 2059090,
      "postDate": "2022-12-08T13:27:40.300Z",
      "content": "<p>Is the kernel assigned with different hardware resources?</p>\n<p>Since the public shared notebooks with LB scores, why my forked learning ones time out?</p>",
      "rawMarkdown": "Is the kernel assigned with different hardware resources?\n\nSince the public shared notebooks with LB scores, why my forked learning ones time out?",
      "votes": 2
    },
    {
      "id": 2060069,
      "postDate": "2022-12-09T14:21:48.453Z",
      "content": "<p>not sure what caused the problem. Tried again, worked.</p>",
      "rawMarkdown": "not sure what caused the problem. Tried again, worked."
    }
  ],
  "comments": [
    {
      "id": 2059311,
      "author_name": "@kaggleqrdl",
      "author_url": "",
      "post_date": "2022-12-08T17:41:47.833000",
      "content": "<p>Which one is timing out?  Agreed, that is problematic for sure.  And yes, different hardware resources is a common issue which has been raised before.</p>\n<p>You could try using the faster dicom decode library, dicomsdl </p>\n<p>Note if Kaggle supplied us with lz4 compressed files of the pixel data instead of dicom, we'd see up to a ~10x speedup in decoding, though it would take about 2x the amount of space.</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 2060069,
      "author_name": "dragon zhang",
      "author_url": "",
      "post_date": "2022-12-09T14:21:48.453000",
      "content": "<p>not sure what caused the problem. Tried again, worked.</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2059311": "Which one is timing out?  Agreed, that is problematic for sure.  And yes, different hardware resources is a common issue which has been raised before.\n\nYou could try using the faster dicom decode library, dicomsdl \n\nNote if Kaggle supplied us with lz4 compressed files of the pixel data instead of dicom, we'd see up to a ~10x speedup in decoding, though it would take about 2x the amount of space.",
    "2059090": "Is the kernel assigned with different hardware resources?\n\nSince the public shared notebooks with LB scores, why my forked learning ones time out?",
    "2060069": "not sure what caused the problem. Tried again, worked."
  }
}