{
  "id": 145364,
  "title": "Downloads are fixed",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/145364",
  "author_name": "Sohier Dane",
  "post_date": "2020-04-22T20:10:23.632000",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Downloading the full public dataset should be working again now, both via the \"download all\" button and the API. </p>\n\n<p>Apologies for the hassle! </p>",
  "messages": [
    {
      "id": 817074,
      "postDate": "2020-04-22T20:10:23.633Z",
      "content": "<p>Downloading the full public dataset should be working again now, both via the \"download all\" button and the API. </p>\n\n<p>Apologies for the hassle! </p>",
      "rawMarkdown": "Downloading the full public dataset should be working again now, both via the \"download all\" button and the API. \n\nApologies for the hassle! ",
      "votes": 3
    },
    {
      "id": 817103,
      "postDate": "2020-04-22T20:47:19.220Z",
      "content": "<p>Yep, it's working, thanks a lot !  </p>\n\n<p>However, to download hundreds of Gb without any incidents might not be easy. On my part, the DL is going to take 24hours. Any internet interruption and I'll have to start again. Wouldn't it be possible somehow to split it in more manageable chunks ? Or a least make it possible to download the train_images and train masks separately ? I suppose it would relieve the kaggle servers as well, for most people won't need the train masks at first, I guess.</p>\n\n<p>Any thoughts ? </p>\n\n<p>Again, Thanks a lot for your work ! </p>",
      "rawMarkdown": "Yep, it's working, thanks a lot !  \n\nHowever, to download hundreds of Gb without any incidents might not be easy. On my part, the DL is going to take 24hours. Any internet interruption and I'll have to start again. Wouldn't it be possible somehow to split it in more manageable chunks ? Or a least make it possible to download the train_images and train masks separately ? I suppose it would relieve the kaggle servers as well, for most people won't need the train masks at first, I guess.\n\nAny thoughts ? \n\nAgain, Thanks a lot for your work ! \n",
      "votes": 2
    },
    {
      "id": 879098,
      "postDate": "2020-06-09T07:33:08.053Z",
      "content": "<p>I am a late entrant with a flaky connection. Can you please provide a wget path?\nOr can you please provide a --continue option in the api?</p>",
      "rawMarkdown": "I am a late entrant with a flaky connection. Can you please provide a wget path?\nOr can you please provide a --continue option in the api?"
    }
  ],
  "comments": [
    {
      "id": 817103,
      "author_name": "Benjamin Dubreu",
      "author_url": "",
      "post_date": "2020-04-22T20:47:19.220000",
      "content": "<p>Yep, it's working, thanks a lot !  </p>\n\n<p>However, to download hundreds of Gb without any incidents might not be easy. On my part, the DL is going to take 24hours. Any internet interruption and I'll have to start again. Wouldn't it be possible somehow to split it in more manageable chunks ? Or a least make it possible to download the train_images and train masks separately ? I suppose it would relieve the kaggle servers as well, for most people won't need the train masks at first, I guess.</p>\n\n<p>Any thoughts ? </p>\n\n<p>Again, Thanks a lot for your work ! </p>",
      "votes": 2,
      "replies": []
    },
    {
      "id": 879098,
      "author_name": "Ashish Lal",
      "author_url": "",
      "post_date": "2020-06-09T07:33:08.053000",
      "content": "<p>I am a late entrant with a flaky connection. Can you please provide a wget path?\nOr can you please provide a --continue option in the api?</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "817074": "Downloading the full public dataset should be working again now, both via the \"download all\" button and the API. \n\nApologies for the hassle! ",
    "817103": "Yep, it's working, thanks a lot !  \n\nHowever, to download hundreds of Gb without any incidents might not be easy. On my part, the DL is going to take 24hours. Any internet interruption and I'll have to start again. Wouldn't it be possible somehow to split it in more manageable chunks ? Or a least make it possible to download the train_images and train masks separately ? I suppose it would relieve the kaggle servers as well, for most people won't need the train masks at first, I guess.\n\nAny thoughts ? \n\nAgain, Thanks a lot for your work ! \n",
    "879098": "I am a late entrant with a flaky connection. Can you please provide a wget path?\nOr can you please provide a --continue option in the api?"
  }
}