{
  "id": 124287,
  "title": "Notebook exceeded allowed compute",
  "url": "/competitions/bengaliai-cv19/discussion/124287",
  "author_name": "Son of Anton v3.0",
  "post_date": "2020-01-03T07:22:28.857000",
  "votes": 2,
  "comment_count": 6,
  "views": 0,
  "content": "<p>I get \"Notebook exceeded allowed compute\" when trying to submit submission.csv.</p>",
  "messages": [
    {
      "id": 709242,
      "postDate": "2020-01-03T08:25:31.693Z",
      "content": "<p>This could be due to time limit; for Kernel Only Competition we need to do inference under <code>2 hours</code>. It includes both <code>public</code> and <code>private</code> test sets. Make sure that your kernel runs under 2 hours </p>",
      "rawMarkdown": "This could be due to time limit; for Kernel Only Competition we need to do inference under `2 hours`. It includes both `public` and `private` test sets. Make sure that your kernel runs under 2 hours ",
      "votes": 1
    },
    {
      "id": 709215,
      "postDate": "2020-01-03T07:22:28.857Z",
      "content": "<p>I get \"Notebook exceeded allowed compute\" when trying to submit submission.csv.</p>",
      "rawMarkdown": "I get \"Notebook exceeded allowed compute\" when trying to submit submission.csv.",
      "votes": 2
    },
    {
      "id": 758414,
      "postDate": "2020-02-27T18:18:51.577Z",
      "content": "<p>Tried almost every optimisations mentioned, still not even a single submission.\nThings i tried : \n1. Inference only kernel ( No training)\n2. Delete every variable after the use (memory optimisations)\n3. Prediction on batches instead of single inference.</p>\n\n<p>One thing I noted, that the cpu goes above 100%.</p>\n\n<p>Can You suggest something.?</p>",
      "rawMarkdown": "Tried almost every optimisations mentioned, still not even a single submission.\nThings i tried : \n1. Inference only kernel ( No training)\n2. Delete every variable after the use (memory optimisations)\n3. Prediction on batches instead of single inference.\n\nOne thing I noted, that the cpu goes above 100%.\n\nCan You suggest something.?\n"
    },
    {
      "id": 716119,
      "postDate": "2020-01-11T09:34:58.687Z",
      "content": "<p>Did you resolved this problem ? , i'm getting the same Error after 5 min run </p>",
      "rawMarkdown": "Did you resolved this problem ? , i'm getting the same Error after 5 min run ",
      "replies": [
        {
          "id": 718171,
          "postDate": "2020-01-14T05:48:53.523Z",
          "content": "<p>I resolved it. Download the model and load it into a separate kernel which you can use it club your predictions and submit. Ensure that this kernel doesn't have any training part. Ensure that you deallocate the variables that you have previously used(Ex.The test parquet files). You can use del </p>",
          "rawMarkdown": "I resolved it. Download the model and load it into a separate kernel which you can use it club your predictions and submit. Ensure that this kernel doesn't have any training part. Ensure that you deallocate the variables that you have previously used(Ex.The test parquet files). You can use del "
        }
      ]
    },
    {
      "id": 711797,
      "postDate": "2020-01-06T14:44:31.750Z",
      "content": "<p>Make sure no training is happening in the kernel.</p>",
      "rawMarkdown": "Make sure no training is happening in the kernel."
    },
    {
      "id": 718170,
      "postDate": "2020-01-14T05:48:11.957Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 709242,
      "author_name": "Bibek",
      "author_url": "",
      "post_date": "2020-01-03T08:25:31.693000",
      "content": "<p>This could be due to time limit; for Kernel Only Competition we need to do inference under <code>2 hours</code>. It includes both <code>public</code> and <code>private</code> test sets. Make sure that your kernel runs under 2 hours </p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 758414,
      "author_name": "Avinash Rai",
      "author_url": "",
      "post_date": "2020-02-27T18:18:51.577000",
      "content": "<p>Tried almost every optimisations mentioned, still not even a single submission.\nThings i tried : \n1. Inference only kernel ( No training)\n2. Delete every variable after the use (memory optimisations)\n3. Prediction on batches instead of single inference.</p>\n\n<p>One thing I noted, that the cpu goes above 100%.</p>\n\n<p>Can You suggest something.?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 716119,
      "author_name": "Shay Mitrani",
      "author_url": "",
      "post_date": "2020-01-11T09:34:58.687000",
      "content": "<p>Did you resolved this problem ? , i'm getting the same Error after 5 min run </p>",
      "votes": 0,
      "replies": [
        {
          "id": 718171,
          "author_name": "Son of Anton v3.0",
          "author_url": "",
          "post_date": "2020-01-14T05:48:53.523000",
          "content": "<p>I resolved it. Download the model and load it into a separate kernel which you can use it club your predictions and submit. Ensure that this kernel doesn't have any training part. Ensure that you deallocate the variables that you have previously used(Ex.The test parquet files). You can use del </p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 711797,
      "author_name": "ShubhankarMohan",
      "author_url": "",
      "post_date": "2020-01-06T14:44:31.750000",
      "content": "<p>Make sure no training is happening in the kernel.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 718170,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-01-14T05:48:11.957000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "709242": "This could be due to time limit; for Kernel Only Competition we need to do inference under `2 hours`. It includes both `public` and `private` test sets. Make sure that your kernel runs under 2 hours ",
    "709215": "I get \"Notebook exceeded allowed compute\" when trying to submit submission.csv.",
    "758414": "Tried almost every optimisations mentioned, still not even a single submission.\nThings i tried : \n1. Inference only kernel ( No training)\n2. Delete every variable after the use (memory optimisations)\n3. Prediction on batches instead of single inference.\n\nOne thing I noted, that the cpu goes above 100%.\n\nCan You suggest something.?\n",
    "716119": "Did you resolved this problem ? , i'm getting the same Error after 5 min run ",
    "711797": "Make sure no training is happening in the kernel.",
    "718170": ""
  }
}