{
  "id": 214392,
  "title": "NB Crashing on inference",
  "url": "/competitions/rfcx-species-audio-detection/discussion/214392",
  "author_name": "",
  "post_date": "2021-01-26T13:50:02.428160400Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>My notebook keeps crashing during inference. I have tried inference with two models. I am not using tfrecords, but am instead using pytorch lightning for the whole process. Is anyone else facing the same kind of issue?</p>",
  "messages": [
    {
      "id": "1170856",
      "postDate": "01/26/2021 13:50:02",
      "content": "<p>My notebook keeps crashing during inference. I have tried inference with two models. I am not using tfrecords, but am instead using pytorch lightning for the whole process. Is anyone else facing the same kind of issue?</p>",
      "rawMarkdown": "My notebook keeps crashing during inference. I have tried inference with two models. I am not using tfrecords, but am instead using pytorch lightning for the whole process. Is anyone else facing the same kind of issue?",
      "votes": null
    },
    {
      "id": "1172760",
      "postDate": "01/27/2021 14:40:00",
      "content": "<p>You don't give any detail, which is probably why you don't get any answer.  I would suspect an out of memory error, i.e. that you are using too much memory.</p>",
      "rawMarkdown": "You don't give any detail, which is probably why you don't get any answer.  I would suspect an out of memory error, i.e. that you are using too much memory.",
      "votes": null
    },
    {
      "id": "1172789",
      "postDate": "01/27/2021 14:54:34",
      "content": "<p>Apologies here's my notebook: <a href=\"https://www.kaggle.com/aryaman1999/transformers-for-audio-an-experiment\" target=\"_blank\">https://www.kaggle.com/aryaman1999/transformers-for-audio-an-experiment</a></p>\n<p>I usually get an error in the last cell, which I have commented currently.</p>\n<p>I understand that I may be using too much memory, but is there a way to check exactly how much? If I was to run it locally I would probably use htop, how do I do that in a Kaggle NB</p>",
      "rawMarkdown": "Apologies here's my notebook: https://www.kaggle.com/aryaman1999/transformers-for-audio-an-experiment\n\nI usually get an error in the last cell, which I have commented currently.\n\nI understand that I may be using too much memory, but is there a way to check exactly how much? If I was to run it locally I would probably use htop, how do I do that in a Kaggle NB",
      "votes": null
    },
    {
      "id": "1173132",
      "postDate": "01/27/2021 17:31:10",
      "content": "<p>execute this in a cell:</p>\n<p><code>! top -o %MEM</code></p>",
      "rawMarkdown": "execute this in a cell:\n\n`! top -o %MEM`",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1172760,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "01/27/2021 14:40:00",
      "content": "<p>You don't give any detail, which is probably why you don't get any answer.  I would suspect an out of memory error, i.e. that you are using too much memory.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1172789,
          "author_name": "aryaman1999",
          "author_url": "",
          "post_date": "01/27/2021 14:54:34",
          "content": "<p>Apologies here's my notebook: <a href=\"https://www.kaggle.com/aryaman1999/transformers-for-audio-an-experiment\" target=\"_blank\">https://www.kaggle.com/aryaman1999/transformers-for-audio-an-experiment</a></p>\n<p>I usually get an error in the last cell, which I have commented currently.</p>\n<p>I understand that I may be using too much memory, but is there a way to check exactly how much? If I was to run it locally I would probably use htop, how do I do that in a Kaggle NB</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1173132,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "01/27/2021 17:31:10",
          "content": "<p>execute this in a cell:</p>\n<p><code>! top -o %MEM</code></p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1170856": "My notebook keeps crashing during inference. I have tried inference with two models. I am not using tfrecords, but am instead using pytorch lightning for the whole process. Is anyone else facing the same kind of issue?",
    "1172760": "You don't give any detail, which is probably why you don't get any answer.  I would suspect an out of memory error, i.e. that you are using too much memory.",
    "1172789": "Apologies here's my notebook: https://www.kaggle.com/aryaman1999/transformers-for-audio-an-experiment\n\nI usually get an error in the last cell, which I have commented currently.\n\nI understand that I may be using too much memory, but is there a way to check exactly how much? If I was to run it locally I would probably use htop, how do I do that in a Kaggle NB",
    "1173132": "execute this in a cell:\n\n`! top -o %MEM`"
  },
  "source": "meta"
}