{
  "id": 38240,
  "title": "is it possible to complete this challenge , without downloading any file on local machine. ",
  "url": "/competitions/carvana-image-masking-challenge/discussion/38240",
  "author_name": "",
  "post_date": "2017-08-17T12:19:30.283158400Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I am new to kaggle kernels. I noticed that it take considerable time to train even baseline models for this and other simpler datasets here on kaggle kernels. so I was wondering, is it possible to analyse train test etc everything using kaggle kernels computational power only. If yes , is there some guide for optimization of the performance or  I just need to keep my pc connected and let it run for hours. </p>",
  "messages": [
    {
      "id": "214569",
      "postDate": "08/17/2017 12:19:30",
      "content": "<p>I am new to kaggle kernels. I noticed that it take considerable time to train even baseline models for this and other simpler datasets here on kaggle kernels. so I was wondering, is it possible to analyse train test etc everything using kaggle kernels computational power only. If yes , is there some guide for optimization of the performance or  I just need to keep my pc connected and let it run for hours. </p>",
      "rawMarkdown": "I am new to kaggle kernels. I noticed that it take considerable time to train even baseline models for this and other simpler datasets here on kaggle kernels. so I was wondering, is it possible to analyse train test etc everything using kaggle kernels computational power only. If yes , is there some guide for optimization of the performance or  I just need to keep my pc connected and let it run for hours.",
      "votes": null
    },
    {
      "id": "214572",
      "postDate": "08/17/2017 12:31:57",
      "content": "<p>I would say no. Even the baseline models (128x128 U-nets) would probably time out training on gaggle kernels.</p>",
      "rawMarkdown": "I would say no. Even the baseline models (128x128 U-nets) would probably time out training on gaggle kernels.",
      "votes": null
    },
    {
      "id": "214664",
      "postDate": "08/17/2017 18:53:56",
      "content": "<p>No. Kaggle kernels run for only a maximum of 20 minutes. So one can hardly run most scripts with large data to completion on Kaggle.</p>",
      "rawMarkdown": "No. Kaggle kernels run for only a maximum of 20 minutes. So one can hardly run most scripts with large data to completion on Kaggle.",
      "votes": null
    },
    {
      "id": "214859",
      "postDate": "08/18/2017 15:22:16",
      "content": "<p>Sorry for no free lunch, as we all know. </p>",
      "rawMarkdown": "Sorry for no free lunch, as we all know.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 214572,
      "author_name": "craigglastonbury",
      "author_url": "",
      "post_date": "08/17/2017 12:31:57",
      "content": "<p>I would say no. Even the baseline models (128x128 U-nets) would probably time out training on gaggle kernels.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 214664,
      "author_name": "sheriytm",
      "author_url": "",
      "post_date": "08/17/2017 18:53:56",
      "content": "<p>No. Kaggle kernels run for only a maximum of 20 minutes. So one can hardly run most scripts with large data to completion on Kaggle.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 214859,
      "author_name": "netphone",
      "author_url": "",
      "post_date": "08/18/2017 15:22:16",
      "content": "<p>Sorry for no free lunch, as we all know. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "214569": "I am new to kaggle kernels. I noticed that it take considerable time to train even baseline models for this and other simpler datasets here on kaggle kernels. so I was wondering, is it possible to analyse train test etc everything using kaggle kernels computational power only. If yes , is there some guide for optimization of the performance or  I just need to keep my pc connected and let it run for hours.",
    "214572": "I would say no. Even the baseline models (128x128 U-nets) would probably time out training on gaggle kernels.",
    "214664": "No. Kaggle kernels run for only a maximum of 20 minutes. So one can hardly run most scripts with large data to completion on Kaggle.",
    "214859": "Sorry for no free lunch, as we all know."
  },
  "source": "meta"
}