{
  "id": 203560,
  "title": "Train data availability during submission",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/203560",
  "author_name": "",
  "post_date": "2020-12-15T18:20:16.218781300Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>is the train data accessible in the inference notebook during submission</p>",
  "messages": [
    {
      "id": "1113848",
      "postDate": "12/15/2020 18:20:16",
      "content": "<p>is the train data accessible in the inference notebook during submission</p>",
      "rawMarkdown": "is the train data accessible in the inference notebook during submission",
      "votes": null
    },
    {
      "id": "1113884",
      "postDate": "12/15/2020 18:55:05",
      "content": "<p>Yes - several shared kernels that go from start to train to submission.  Not sure I recall a competition where the train data was no longer available.  The thing that's not as available as needed is time.  Getting a model trained and making predictions often hard to do in the time limit.</p>",
      "rawMarkdown": "Yes - several shared kernels that go from start to train to submission.  Not sure I recall a competition where the train data was no longer available.  The thing that's not as available as needed is time.  Getting a model trained and making predictions often hard to do in the time limit.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1113884,
      "author_name": "pcjimmmy",
      "author_url": "",
      "post_date": "12/15/2020 18:55:05",
      "content": "<p>Yes - several shared kernels that go from start to train to submission.  Not sure I recall a competition where the train data was no longer available.  The thing that's not as available as needed is time.  Getting a model trained and making predictions often hard to do in the time limit.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1113848": "is the train data accessible in the inference notebook during submission",
    "1113884": "Yes - several shared kernels that go from start to train to submission.  Not sure I recall a competition where the train data was no longer available.  The thing that's not as available as needed is time.  Getting a model trained and making predictions often hard to do in the time limit."
  },
  "source": "meta"
}