{
  "id": 340791,
  "title": "Requirements for using TPU? [Long queue]",
  "url": "/competitions/AI4Code/discussion/340791",
  "author_name": "",
  "post_date": "2022-07-31T03:30:36.491858Z",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Good day Kaggle community. I am fairly new to the platform, so would like some advice.<br>\nIn this competition since most language models are fairly large, I was thinking this would be a chance for me to learn to use TPU. Previous week, I was running my notebook just fine. However, since the TPU reset, I was not able to run any TPU notebook, they are all put into Queued state. I have waited as long as half a day until I gave up.<br>\nIs this just a problem for me or is everyone experiencing the same high queue time? If it is just me is there maybe some requirements that needed to be achieve before using more TPU? My profile still show 20h available for TPU, so I wouldn't think I have used up any resource yet.</p>",
  "messages": [
    {
      "id": "1877847",
      "postDate": "07/31/2022 03:30:36",
      "content": "<p>Good day Kaggle community. I am fairly new to the platform, so would like some advice.<br>\nIn this competition since most language models are fairly large, I was thinking this would be a chance for me to learn to use TPU. Previous week, I was running my notebook just fine. However, since the TPU reset, I was not able to run any TPU notebook, they are all put into Queued state. I have waited as long as half a day until I gave up.<br>\nIs this just a problem for me or is everyone experiencing the same high queue time? If it is just me is there maybe some requirements that needed to be achieve before using more TPU? My profile still show 20h available for TPU, so I wouldn't think I have used up any resource yet.</p>",
      "rawMarkdown": "Good day Kaggle community. I am fairly new to the platform, so would like some advice.\nIn this competition since most language models are fairly large, I was thinking this would be a chance for me to learn to use TPU. Previous week, I was running my notebook just fine. However, since the TPU reset, I was not able to run any TPU notebook, they are all put into Queued state. I have waited as long as half a day until I gave up.\nIs this just a problem for me or is everyone experiencing the same high queue time? If it is just me is there maybe some requirements that needed to be achieve before using more TPU? My profile still show 20h available for TPU, so I wouldn't think I have used up any resource yet.",
      "votes": null
    },
    {
      "id": "1877940",
      "postDate": "07/31/2022 04:50:28",
      "content": "<p>This largely depends on the overall demand and supply equation. You could perhaps wait for your chance or maybe run the relevant code elsewhere (colab pro/ other cloud platforms offering GPU rental services) temporarily till this is stabilized.</p>",
      "rawMarkdown": "This largely depends on the overall demand and supply equation. You could perhaps wait for your chance or maybe run the relevant code elsewhere (colab pro/ other cloud platforms offering GPU rental services) temporarily till this is stabilized.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1877940,
      "author_name": "ravi20076",
      "author_url": "",
      "post_date": "07/31/2022 04:50:28",
      "content": "<p>This largely depends on the overall demand and supply equation. You could perhaps wait for your chance or maybe run the relevant code elsewhere (colab pro/ other cloud platforms offering GPU rental services) temporarily till this is stabilized.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1877847": "Good day Kaggle community. I am fairly new to the platform, so would like some advice.\nIn this competition since most language models are fairly large, I was thinking this would be a chance for me to learn to use TPU. Previous week, I was running my notebook just fine. However, since the TPU reset, I was not able to run any TPU notebook, they are all put into Queued state. I have waited as long as half a day until I gave up.\nIs this just a problem for me or is everyone experiencing the same high queue time? If it is just me is there maybe some requirements that needed to be achieve before using more TPU? My profile still show 20h available for TPU, so I wouldn't think I have used up any resource yet.",
    "1877940": "This largely depends on the overall demand and supply equation. You could perhaps wait for your chance or maybe run the relevant code elsewhere (colab pro/ other cloud platforms offering GPU rental services) temporarily till this is stabilized."
  },
  "source": "meta"
}