{
  "id": 103652,
  "title": "Submission: Loading Model versus Training Model",
  "url": "/competitions/aptos2019-blindness-detection/discussion/103652",
  "author_name": "",
  "post_date": "2019-08-10T16:06:01.315359800Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Do we have to train our model in the code that we submit? <br>\nOr can we train our model offline and upload the trained model so the pseudo-code for our submission would be the following?</p>\n\n<p>read test.csv\nread corresponding images\nmy_predictor = get previously trained model\nget results of sending images through my_predictor\noutput submission.csv</p>",
  "messages": [
    {
      "id": "596418",
      "postDate": "08/10/2019 16:06:01",
      "content": "<p>Do we have to train our model in the code that we submit? <br>\nOr can we train our model offline and upload the trained model so the pseudo-code for our submission would be the following?</p>\n\n<p>read test.csv\nread corresponding images\nmy_predictor = get previously trained model\nget results of sending images through my_predictor\noutput submission.csv</p>",
      "rawMarkdown": "Do we have to train our model in the code that we submit?  \nOr can we train our model offline and upload the trained model so the pseudo-code for our submission would be the following?\n\nread test.csv\nread corresponding images\nmy_predictor = get previously trained model\nget results of sending images through my_predictor\noutput submission.csv",
      "votes": null
    },
    {
      "id": "597406",
      "postDate": "08/12/2019 09:41:41",
      "content": "<p>both ok, depends on you.</p>",
      "rawMarkdown": "both ok, depends on you.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 597406,
      "author_name": "frank518",
      "author_url": "",
      "post_date": "08/12/2019 09:41:41",
      "content": "<p>both ok, depends on you.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "596418": "Do we have to train our model in the code that we submit?  \nOr can we train our model offline and upload the trained model so the pseudo-code for our submission would be the following?\n\nread test.csv\nread corresponding images\nmy_predictor = get previously trained model\nget results of sending images through my_predictor\noutput submission.csv",
    "597406": "both ok, depends on you."
  },
  "source": "meta"
}