{
  "id": 151088,
  "title": "CPU Ram, GPU and TPU memory limits",
  "url": "/competitions/jigsaw-multilingual-toxic-comment-classification/discussion/151088",
  "author_name": "",
  "post_date": "2020-05-14T09:44:15.978173800Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Can we get an understanding of the CPU limits, GPU and TPU ram available to the participants? It seems like I am running to memory issues quite often lately. Is the cpu ram limit at 16GB? We are having problem innovating with larger models if the CPU/GPU/TPU ram limit is of current setting. Can Kaggle consider increasing the limit slightly so that model slightly larger than roberta can be load? Many thanks. </p>",
  "messages": [
    {
      "id": "847307",
      "postDate": "05/14/2020 09:44:15",
      "content": "<p>Can we get an understanding of the CPU limits, GPU and TPU ram available to the participants? It seems like I am running to memory issues quite often lately. Is the cpu ram limit at 16GB? We are having problem innovating with larger models if the CPU/GPU/TPU ram limit is of current setting. Can Kaggle consider increasing the limit slightly so that model slightly larger than roberta can be load? Many thanks. </p>",
      "rawMarkdown": "Can we get an understanding of the CPU limits, GPU and TPU ram available to the participants? It seems like I am running to memory issues quite often lately. Is the cpu ram limit at 16GB? We are having problem innovating with larger models if the CPU/GPU/TPU ram limit is of current setting. Can Kaggle consider increasing the limit slightly so that model slightly larger than roberta can be load? Many thanks.",
      "votes": null
    },
    {
      "id": "847310",
      "postDate": "05/14/2020 09:46:56",
      "content": "<p>When running large datasets, it is often discouraging to hit memory limits. With Colab Pro, a user can get priority access to high-memory VMs, which have twice the memory. In other words, Colab users can use up to 12 GB of memory, while Pro users can enjoy up to 25 GB of memory up to availability.Mar 15, 2020,  Anyway to let participants have that limit up to 25GB for this competition? Thks</p>",
      "rawMarkdown": "When running large datasets, it is often discouraging to hit memory limits. With Colab Pro, a user can get priority access to high-memory VMs, which have twice the memory. In other words, Colab users can use up to 12 GB of memory, while Pro users can enjoy up to 25 GB of memory up to availability.Mar 15, 2020,  Anyway to let participants have that limit up to 25GB for this competition? Thks",
      "votes": null
    },
    {
      "id": "853135",
      "postDate": "05/19/2020 01:49:49",
      "content": "<p>When GPU/TPU are turned off ,the CPU RAM limit is 16GB and you have 4 cores. When GPU/TPU are turned on the CPU RAM limit is 13GB and you have 2 cores</p>",
      "rawMarkdown": "When GPU/TPU are turned off ,the CPU RAM limit is 16GB and you have 4 cores. When GPU/TPU are turned on the CPU RAM limit is 13GB and you have 2 cores",
      "votes": null
    },
    {
      "id": "1216806",
      "postDate": "02/24/2021 13:48:51",
      "content": "<p>You can use 30 hours per week of TPUs and up to 9h at a time in a single session. <a href=\"https://www.kaggle.com/drpatrickchan\" target=\"_blank\">@drpatrickchan</a> </p>",
      "rawMarkdown": "You can use 30 hours per week of TPUs and up to 9h at a time in a single session. @drpatrickchan",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1216806,
      "author_name": "riyadhar",
      "author_url": "",
      "post_date": "02/24/2021 13:48:51",
      "content": "<p>You can use 30 hours per week of TPUs and up to 9h at a time in a single session. <a href=\"https://www.kaggle.com/drpatrickchan\" target=\"_blank\">@drpatrickchan</a> </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 847310,
      "author_name": "drpatrickchan",
      "author_url": "",
      "post_date": "05/14/2020 09:46:56",
      "content": "<p>When running large datasets, it is often discouraging to hit memory limits. With Colab Pro, a user can get priority access to high-memory VMs, which have twice the memory. In other words, Colab users can use up to 12 GB of memory, while Pro users can enjoy up to 25 GB of memory up to availability.Mar 15, 2020,  Anyway to let participants have that limit up to 25GB for this competition? Thks</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 853135,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "05/19/2020 01:49:49",
      "content": "<p>When GPU/TPU are turned off ,the CPU RAM limit is 16GB and you have 4 cores. When GPU/TPU are turned on the CPU RAM limit is 13GB and you have 2 cores</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "847307": "Can we get an understanding of the CPU limits, GPU and TPU ram available to the participants? It seems like I am running to memory issues quite often lately. Is the cpu ram limit at 16GB? We are having problem innovating with larger models if the CPU/GPU/TPU ram limit is of current setting. Can Kaggle consider increasing the limit slightly so that model slightly larger than roberta can be load? Many thanks.",
    "847310": "When running large datasets, it is often discouraging to hit memory limits. With Colab Pro, a user can get priority access to high-memory VMs, which have twice the memory. In other words, Colab users can use up to 12 GB of memory, while Pro users can enjoy up to 25 GB of memory up to availability.Mar 15, 2020,  Anyway to let participants have that limit up to 25GB for this competition? Thks",
    "853135": "When GPU/TPU are turned off ,the CPU RAM limit is 16GB and you have 4 cores. When GPU/TPU are turned on the CPU RAM limit is 13GB and you have 2 cores",
    "1216806": "You can use 30 hours per week of TPUs and up to 9h at a time in a single session. @drpatrickchan"
  },
  "source": "meta"
}