{
  "id": 116032,
  "title": "Any other surprises? Why train images removed!?",
  "url": "/competitions/rsna-intracranial-hemorrhage-detection/discussion/116032",
  "author_name": "Mukharbek Organokov",
  "post_date": "2019-11-06T17:28:45.908000",
  "votes": 10,
  "comment_count": 8,
  "views": 0,
  "content": "<p>How we can train our models on GCP if the train images removed. We didn't download that. \nI, personally, have no storage space and didn't in advance download it to my laptop. </p>\n\n<p>sad(</p>",
  "messages": [
    {
      "id": 666986,
      "postDate": "2019-11-06T17:28:45.910Z",
      "content": "<p>How we can train our models on GCP if the train images removed. We didn't download that. \nI, personally, have no storage space and didn't in advance download it to my laptop. </p>\n\n<p>sad(</p>",
      "rawMarkdown": "How we can train our models on GCP if the train images removed. We didn't download that. \nI, personally, have no storage space and didn't in advance download it to my laptop. \n\nsad(",
      "votes": 10
    },
    {
      "id": 666988,
      "postDate": "2019-11-06T17:31:26.547Z",
      "content": "<p><a href=\"/juliaelliott\">@juliaelliott</a> <a href=\"/philculliton\">@philculliton</a> </p>",
      "rawMarkdown": "@juliaelliott @philculliton \n",
      "votes": 4,
      "replies": [
        {
          "id": 667300,
          "postDate": "2019-11-07T05:04:48.373Z",
          "content": "<p>12 hours gone,,no reply,,thanks to this competition</p>",
          "rawMarkdown": "12 hours gone,,no reply,,thanks to this competition",
          "votes": 2
        }
      ]
    },
    {
      "id": 666992,
      "postDate": "2019-11-06T17:38:19.230Z",
      "content": "<p>Did you train on GCP for stage 1?</p>",
      "rawMarkdown": "Did you train on GCP for stage 1?",
      "votes": 1,
      "replies": [
        {
          "id": 666994,
          "postDate": "2019-11-06T17:42:04.583Z",
          "content": "<p><a href=\"/cparrarojas\">@cparrarojas</a> </p>\n\n<p>We uploaded 2 models: one trained on GCP, another one with kernel only\nand that from GCP was better on public LB and we don't want to lose it because for some reason it was decided sadly to close access to data directly. </p>",
          "rawMarkdown": "@cparrarojas \n\nWe uploaded 2 models: one trained on GCP, another one with kernel only\nand that from GCP was better on public LB and we don't want to lose it because for some reason it was decided sadly to close access to data directly. ",
          "votes": 2
        },
        {
          "id": 667002,
          "postDate": "2019-11-06T17:50:50.060Z",
          "content": "<p>So you have the stage 1 data on your GCP instance then, no?</p>",
          "rawMarkdown": "So you have the stage 1 data on your GCP instance then, no?",
          "votes": 1
        },
        {
          "id": 667011,
          "postDate": "2019-11-06T18:02:30.863Z",
          "content": "<p>Nope it was actually taking lot of space. So had to remove it from GCP instance. </p>",
          "rawMarkdown": "Nope it was actually taking lot of space. So had to remove it from GCP instance. ",
          "votes": 1
        }
      ]
    },
    {
      "id": 667000,
      "postDate": "2019-11-06T17:50:03.010Z",
      "content": "<p><a href=\"/mobassir\">@mobassir</a>  <a href=\"/manojprabhaakr\">@manojprabhaakr</a> <a href=\"/carlolepelaars\">@carlolepelaars</a> \nWe can have test images from <a href=\"/hengck23\">@hengck23</a> by this link\n<a href=\"https://drive.google.com/open?id=16P7OizMu8i7PzRZhNkzk_I4HgwtGOrYM\">https://drive.google.com/open?id=16P7OizMu8i7PzRZhNkzk_I4HgwtGOrYM</a>\nas posted in his topic\n<a href=\"https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/115855\">https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/115855</a></p>\n\n<p>now we need to find 180+ Gb space and the link for train images</p>",
      "rawMarkdown": "@mobassir  @manojprabhaakr @carlolepelaars \nWe can have test images from @hengck23 by this link\nhttps://drive.google.com/open?id=16P7OizMu8i7PzRZhNkzk_I4HgwtGOrYM\nas posted in his topic\nhttps://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/115855\n\nnow we need to find 180+ Gb space and the link for train images",
      "votes": 2,
      "replies": [
        {
          "id": 667299,
          "postDate": "2019-11-07T05:04:07.790Z",
          "content": "<p>test set is here : <a href=\"https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/data\">https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/data</a>\nwe need train set</p>",
          "rawMarkdown": "test set is here : https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/data\nwe need train set",
          "votes": 1
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 666988,
      "author_name": "Mukharbek Organokov",
      "author_url": "",
      "post_date": "2019-11-06T17:31:26.547000",
      "content": "<p><a href=\"/juliaelliott\">@juliaelliott</a> <a href=\"/philculliton\">@philculliton</a> </p>",
      "votes": 4,
      "replies": [
        {
          "id": 667300,
          "author_name": "Mobassir",
          "author_url": "",
          "post_date": "2019-11-07T05:04:48.373000",
          "content": "<p>12 hours gone,,no reply,,thanks to this competition</p>",
          "votes": 2,
          "replies": []
        }
      ]
    },
    {
      "id": 666992,
      "author_name": "César Parra Rojas",
      "author_url": "",
      "post_date": "2019-11-06T17:38:19.230000",
      "content": "<p>Did you train on GCP for stage 1?</p>",
      "votes": 1,
      "replies": [
        {
          "id": 666994,
          "author_name": "Mukharbek Organokov",
          "author_url": "",
          "post_date": "2019-11-06T17:42:04.583000",
          "content": "<p><a href=\"/cparrarojas\">@cparrarojas</a> </p>\n\n<p>We uploaded 2 models: one trained on GCP, another one with kernel only\nand that from GCP was better on public LB and we don't want to lose it because for some reason it was decided sadly to close access to data directly. </p>",
          "votes": 2,
          "replies": []
        },
        {
          "id": 667002,
          "author_name": "César Parra Rojas",
          "author_url": "",
          "post_date": "2019-11-06T17:50:50.060000",
          "content": "<p>So you have the stage 1 data on your GCP instance then, no?</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 667011,
          "author_name": "Manoj Prabhakar",
          "author_url": "",
          "post_date": "2019-11-06T18:02:30.863000",
          "content": "<p>Nope it was actually taking lot of space. So had to remove it from GCP instance. </p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 667000,
      "author_name": "Mukharbek Organokov",
      "author_url": "",
      "post_date": "2019-11-06T17:50:03.010000",
      "content": "<p><a href=\"/mobassir\">@mobassir</a>  <a href=\"/manojprabhaakr\">@manojprabhaakr</a> <a href=\"/carlolepelaars\">@carlolepelaars</a> \nWe can have test images from <a href=\"/hengck23\">@hengck23</a> by this link\n<a href=\"https://drive.google.com/open?id=16P7OizMu8i7PzRZhNkzk_I4HgwtGOrYM\">https://drive.google.com/open?id=16P7OizMu8i7PzRZhNkzk_I4HgwtGOrYM</a>\nas posted in his topic\n<a href=\"https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/115855\">https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/115855</a></p>\n\n<p>now we need to find 180+ Gb space and the link for train images</p>",
      "votes": 2,
      "replies": [
        {
          "id": 667299,
          "author_name": "Mobassir",
          "author_url": "",
          "post_date": "2019-11-07T05:04:07.790000",
          "content": "<p>test set is here : <a href=\"https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/data\">https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/data</a>\nwe need train set</p>",
          "votes": 1,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "666986": "How we can train our models on GCP if the train images removed. We didn't download that. \nI, personally, have no storage space and didn't in advance download it to my laptop. \n\nsad(",
    "666988": "@juliaelliott @philculliton \n",
    "666992": "Did you train on GCP for stage 1?",
    "667000": "@mobassir  @manojprabhaakr @carlolepelaars \nWe can have test images from @hengck23 by this link\nhttps://drive.google.com/open?id=16P7OizMu8i7PzRZhNkzk_I4HgwtGOrYM\nas posted in his topic\nhttps://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/115855\n\nnow we need to find 180+ Gb space and the link for train images"
  }
}