{
  "id": 183289,
  "title": "Submission without generating output",
  "url": "/competitions/landmark-recognition-2020/discussion/183289",
  "author_name": "",
  "post_date": "2020-09-16T06:24:40.500179400Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I have to use GPUs to run my kernel for submission. But I generally exhaust my quota very early in the week.</p>\n<p>As we need to submit submission.csv file along with our code, we need to run our kernel at any cost I guess.</p>\n<p>Is there any alternate way to submit without commiting kernel(generating output)?</p>",
  "messages": [
    {
      "id": "1012530",
      "postDate": "09/16/2020 06:24:40",
      "content": "<p>I have to use GPUs to run my kernel for submission. But I generally exhaust my quota very early in the week.</p>\n<p>As we need to submit submission.csv file along with our code, we need to run our kernel at any cost I guess.</p>\n<p>Is there any alternate way to submit without commiting kernel(generating output)?</p>",
      "rawMarkdown": "I have to use GPUs to run my kernel for submission. But I generally exhaust my quota very early in the week.\n\nAs we need to submit submission.csv file along with our code, we need to run our kernel at any cost I guess.\n\nIs there any alternate way to submit without commiting kernel(generating output)?",
      "votes": null
    },
    {
      "id": "1012769",
      "postDate": "09/16/2020 09:24:25",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/rohitdeepu17\" target=\"_blank\">@rohitdeepu17</a> ,</p>\n<p>You can run the code in google colab and submit the  submission.csv genrated.</p>\n<p>Hope it helps!</p>",
      "rawMarkdown": "Hi @rohitdeepu17 ,\n\nYou can run the code in google colab and submit the  submission.csv genrated.\n\nHope it helps!",
      "votes": null
    },
    {
      "id": "1013645",
      "postDate": "09/16/2020 21:02:18",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/rohitdeepu17\" target=\"_blank\">@rohitdeepu17</a> <a href=\"https://www.kaggle.com/namanbhargava2702\" target=\"_blank\">@namanbhargava2702</a> the mentioned solution is only partially correct.<br>\nYes you can perfectly train your models in Google colab but you can't generate the submission.csv there.</p>\n<p>You need to upload your trained model(s) and then perform the inference in the Kaggle kernels.</p>",
      "rawMarkdown": "Hi @rohitdeepu17 @namanbhargava2702 the mentioned solution is only partially correct.\nYes you can perfectly train your models in Google colab but you can't generate the submission.csv there.\n\nYou need to upload your trained model(s) and then perform the inference in the Kaggle kernels.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1012769,
      "author_name": "namanbhargava2702",
      "author_url": "",
      "post_date": "09/16/2020 09:24:25",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/rohitdeepu17\" target=\"_blank\">@rohitdeepu17</a> ,</p>\n<p>You can run the code in google colab and submit the  submission.csv genrated.</p>\n<p>Hope it helps!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1013645,
          "author_name": "rsmits",
          "author_url": "",
          "post_date": "09/16/2020 21:02:18",
          "content": "<p>Hi <a href=\"https://www.kaggle.com/rohitdeepu17\" target=\"_blank\">@rohitdeepu17</a> <a href=\"https://www.kaggle.com/namanbhargava2702\" target=\"_blank\">@namanbhargava2702</a> the mentioned solution is only partially correct.<br>\nYes you can perfectly train your models in Google colab but you can't generate the submission.csv there.</p>\n<p>You need to upload your trained model(s) and then perform the inference in the Kaggle kernels.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1012530": "I have to use GPUs to run my kernel for submission. But I generally exhaust my quota very early in the week.\n\nAs we need to submit submission.csv file along with our code, we need to run our kernel at any cost I guess.\n\nIs there any alternate way to submit without commiting kernel(generating output)?",
    "1012769": "Hi @rohitdeepu17 ,\n\nYou can run the code in google colab and submit the  submission.csv genrated.\n\nHope it helps!",
    "1013645": "Hi @rohitdeepu17 @namanbhargava2702 the mentioned solution is only partially correct.\nYes you can perfectly train your models in Google colab but you can't generate the submission.csv there.\n\nYou need to upload your trained model(s) and then perform the inference in the Kaggle kernels."
  },
  "source": "meta"
}