{
  "id": 268371,
  "title": "\"Notebook Exceeded Allowed Compute\"",
  "url": "/competitions/landmark-recognition-2021/discussion/268371",
  "author_name": "",
  "post_date": "2021-08-27T05:43:18.455349100Z",
  "votes": 6,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I am getting \"Notebook Exceeded Allowed Compute\" error a lot recently. Currently when we monitor the RAM usage during the run time on the test set it peaks at 9GB and we are still getting the \"Notebook Exceeded Allowed Compute\" error.<br>\nIs anyone else experiencing the same? Is there any solution？</p>",
  "messages": [
    {
      "id": "1492358",
      "postDate": "08/27/2021 05:43:18",
      "content": "<p>I am getting \"Notebook Exceeded Allowed Compute\" error a lot recently. Currently when we monitor the RAM usage during the run time on the test set it peaks at 9GB and we are still getting the \"Notebook Exceeded Allowed Compute\" error.<br>\nIs anyone else experiencing the same? Is there any solution？</p>",
      "rawMarkdown": "I am getting \"Notebook Exceeded Allowed Compute\" error a lot recently. Currently when we monitor the RAM usage during the run time on the test set it peaks at 9GB and we are still getting the \"Notebook Exceeded Allowed Compute\" error.\nIs anyone else experiencing the same? Is there any solution？",
      "votes": null
    },
    {
      "id": "1493118",
      "postDate": "08/27/2021 16:16:51",
      "content": "<p>We usually experience this issue <a href=\"https://www.kaggle.com/lueric\" target=\"_blank\">@lueric</a> when the dataset is huge, in such cases it is better to build the model and save it, then utilize the same in another notebook.</p>",
      "rawMarkdown": "We usually experience this issue @lueric when the dataset is huge, in such cases it is better to build the model and save it, then utilize the same in another notebook.",
      "votes": null
    },
    {
      "id": "1493576",
      "postDate": "08/28/2021 03:37:53",
      "content": "<p>Thanks! In my notebook, I didn't train model, but just loaded my model file, tested the test set, and finally saved the prediction results and generated submission.csv. Your method is to save the prediction results, and then load the prediction results in another notebook and generate submission.csv?</p>",
      "rawMarkdown": "Thanks! In my notebook, I didn't train model, but just loaded my model file, tested the test set, and finally saved the prediction results and generated submission.csv. Your method is to save the prediction results, and then load the prediction results in another notebook and generate submission.csv?",
      "votes": null
    },
    {
      "id": "1496093",
      "postDate": "08/30/2021 05:20:59",
      "content": "<p>Yes, you can try that <a href=\"https://www.kaggle.com/lueric\" target=\"_blank\">@lueric</a> , it should help.</p>",
      "rawMarkdown": "Yes, you can try that @lueric , it should help.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1493118,
      "author_name": "saurabhbagchi",
      "author_url": "",
      "post_date": "08/27/2021 16:16:51",
      "content": "<p>We usually experience this issue <a href=\"https://www.kaggle.com/lueric\" target=\"_blank\">@lueric</a> when the dataset is huge, in such cases it is better to build the model and save it, then utilize the same in another notebook.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1493576,
      "author_name": "lueric",
      "author_url": "",
      "post_date": "08/28/2021 03:37:53",
      "content": "<p>Thanks! In my notebook, I didn't train model, but just loaded my model file, tested the test set, and finally saved the prediction results and generated submission.csv. Your method is to save the prediction results, and then load the prediction results in another notebook and generate submission.csv?</p>",
      "votes": null,
      "replies": [
        {
          "id": 1496093,
          "author_name": "saurabhbagchi",
          "author_url": "",
          "post_date": "08/30/2021 05:20:59",
          "content": "<p>Yes, you can try that <a href=\"https://www.kaggle.com/lueric\" target=\"_blank\">@lueric</a> , it should help.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1492358": "I am getting \"Notebook Exceeded Allowed Compute\" error a lot recently. Currently when we monitor the RAM usage during the run time on the test set it peaks at 9GB and we are still getting the \"Notebook Exceeded Allowed Compute\" error.\nIs anyone else experiencing the same? Is there any solution？",
    "1493118": "We usually experience this issue @lueric when the dataset is huge, in such cases it is better to build the model and save it, then utilize the same in another notebook.",
    "1493576": "Thanks! In my notebook, I didn't train model, but just loaded my model file, tested the test set, and finally saved the prediction results and generated submission.csv. Your method is to save the prediction results, and then load the prediction results in another notebook and generate submission.csv?",
    "1496093": "Yes, you can try that @lueric , it should help."
  },
  "source": "meta"
}