{
  "id": 241399,
  "title": "Would Kaggle change to save partial result instead of cancelling all after timeout?",
  "url": "/competitions/bms-molecular-translation/discussion/241399",
  "author_name": "",
  "post_date": "2021-05-24T12:00:09.955467100Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Today I committed my notebook, because I ignored the preparation and post process, It was cancelled after 9 hours. Actually perhaps another half hour or so, It would be done. Also the output weights files are of values for continuing training.  </p>\n<p>how about save output files in /kaggle/working/  every hour or two?</p>",
  "messages": [
    {
      "id": "1320898",
      "postDate": "05/24/2021 12:00:09",
      "content": "<p>Today I committed my notebook, because I ignored the preparation and post process, It was cancelled after 9 hours. Actually perhaps another half hour or so, It would be done. Also the output weights files are of values for continuing training.  </p>\n<p>how about save output files in /kaggle/working/  every hour or two?</p>",
      "rawMarkdown": "Today I committed my notebook, because I ignored the preparation and post process, It was cancelled after 9 hours. Actually perhaps another half hour or so, It would be done. Also the output weights files are of values for continuing training.  \n\n how about save output files in /kaggle/working/  every hour or two?",
      "votes": null
    },
    {
      "id": "1320940",
      "postDate": "05/24/2021 12:36:36",
      "content": "<p>You can save weights for every k steps (within each epoch) .  They will be there even after notebook timeout</p>",
      "rawMarkdown": "You can save weights for every k steps (within each epoch) .  They will be there even after notebook timeout",
      "votes": null
    },
    {
      "id": "1321020",
      "postDate": "05/24/2021 13:34:15",
      "content": "<p>Thank you!  I don't know that before. </p>",
      "rawMarkdown": "Thank you!  I don't know that before.",
      "votes": null
    },
    {
      "id": "1321495",
      "postDate": "05/24/2021 17:58:52",
      "content": "<p>You will need to save them as dataset if you want to continue training on another notebook. </p>",
      "rawMarkdown": "You will need to save them as dataset if you want to continue training on another notebook.",
      "votes": null
    },
    {
      "id": "1321630",
      "postDate": "05/24/2021 20:08:47",
      "content": "<p>Thanks again. I know that.  Just I almost never check result of my cancelled notebook.. </p>",
      "rawMarkdown": "Thanks again. I know that.  Just I almost never check result of my cancelled notebook..",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1320940,
      "author_name": "serigne",
      "author_url": "",
      "post_date": "05/24/2021 12:36:36",
      "content": "<p>You can save weights for every k steps (within each epoch) .  They will be there even after notebook timeout</p>",
      "votes": null,
      "replies": [
        {
          "id": 1321020,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "05/24/2021 13:34:15",
          "content": "<p>Thank you!  I don't know that before. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1321495,
          "author_name": "serigne",
          "author_url": "",
          "post_date": "05/24/2021 17:58:52",
          "content": "<p>You will need to save them as dataset if you want to continue training on another notebook. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1321630,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "05/24/2021 20:08:47",
          "content": "<p>Thanks again. I know that.  Just I almost never check result of my cancelled notebook.. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1320898": "Today I committed my notebook, because I ignored the preparation and post process, It was cancelled after 9 hours. Actually perhaps another half hour or so, It would be done. Also the output weights files are of values for continuing training.  \n\n how about save output files in /kaggle/working/  every hour or two?",
    "1320940": "You can save weights for every k steps (within each epoch) .  They will be there even after notebook timeout",
    "1321020": "Thank you!  I don't know that before.",
    "1321495": "You will need to save them as dataset if you want to continue training on another notebook.",
    "1321630": "Thanks again. I know that.  Just I almost never check result of my cancelled notebook.."
  },
  "source": "meta"
}