{
  "id": 250734,
  "title": "Submission Error! Notebook Exceeded Allowed Compute! Help Please",
  "url": "/competitions/mlb-player-digital-engagement-forecasting/discussion/250734",
  "author_name": "",
  "post_date": "2021-07-04T07:13:50.370059200Z",
  "votes": 1,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Hi,</p>\n<p>The code seems to run fine on the Kaggle end. After successful commit, If I try to submit it to the competition I get the error \"Notebook Exceeded Allowed Compute\". Can someone help me out figure what is going wrong? </p>\n<p>Thanks in advance.</p>\n<p>My Notebook link:<br>\n<a href=\"url\" target=\"_blank\">https://www.kaggle.com/garggirish/mlb-basic-model-1</a></p>\n<p>I've build a very basic model with randomly chosen features. The idea is that first I want to build a very basic model and work my way upwards from there.</p>",
  "messages": [
    {
      "id": "1375389",
      "postDate": "07/04/2021 07:13:50",
      "content": "<p>Hi,</p>\n<p>The code seems to run fine on the Kaggle end. After successful commit, If I try to submit it to the competition I get the error \"Notebook Exceeded Allowed Compute\". Can someone help me out figure what is going wrong? </p>\n<p>Thanks in advance.</p>\n<p>My Notebook link:<br>\n<a href=\"url\" target=\"_blank\">https://www.kaggle.com/garggirish/mlb-basic-model-1</a></p>\n<p>I've build a very basic model with randomly chosen features. The idea is that first I want to build a very basic model and work my way upwards from there.</p>",
      "rawMarkdown": "Hi,\n\n The code seems to run fine on the Kaggle end. After successful commit, If I try to submit it to the competition I get the error \"Notebook Exceeded Allowed Compute\". Can someone help me out figure what is going wrong? \n\nThanks in advance.\n\nMy Notebook link:\n[https://www.kaggle.com/garggirish/mlb-basic-model-1](url)\n\nI've build a very basic model with randomly chosen features. The idea is that first I want to build a very basic model and work my way upwards from there.",
      "votes": null
    },
    {
      "id": "1376490",
      "postDate": "07/05/2021 06:54:16",
      "content": "<p>By the way the link does not work</p>",
      "rawMarkdown": "By the way the link does not work",
      "votes": null
    },
    {
      "id": "1377120",
      "postDate": "07/05/2021 15:37:16",
      "content": "<p>Can you try copy pasting the link in the browser? I think it opens that way</p>",
      "rawMarkdown": "Can you try copy pasting the link in the browser? I think it opens that way",
      "votes": null
    },
    {
      "id": "1377475",
      "postDate": "07/05/2021 22:02:27",
      "content": "<p>You have to keep an eye on your memory usage and maybe it will help you to solve your problem.</p>",
      "rawMarkdown": "You have to keep an eye on your memory usage and maybe it will help you to solve your problem.",
      "votes": null
    },
    {
      "id": "1379861",
      "postDate": "07/07/2021 16:42:22",
      "content": "<p>Hi, you can pretrain your neural net in the same notebook and save your weights via checkpoints.<br>\n<a href=\"https://www.tensorflow.org/guide/checkpoint\" target=\"_blank\">https://www.tensorflow.org/guide/checkpoint</a></p>\n<p>Then, make next version of your notebook, where you just load weights in the same architecture of net and use pretrained net for submission. So, you will not waste your RAM for training net in submission.</p>",
      "rawMarkdown": "Hi, you can pretrain your neural net in the same notebook and save your weights via checkpoints.\nhttps://www.tensorflow.org/guide/checkpoint\n\nThen, make next version of your notebook, where you just load weights in the same architecture of net and use pretrained net for submission. So, you will not waste your RAM for training net in submission.",
      "votes": null
    },
    {
      "id": "1380321",
      "postDate": "07/08/2021 03:00:43",
      "content": "<p>A problem with this approach - you will only be able to train your models up to the July 20th update of the training files - so you might miss at a minimum 10 days of information, perhaps more if the 7/20 only includes June.  </p>",
      "rawMarkdown": "A problem with this approach - you will only be able to train your models up to the July 20th update of the training files - so you might miss at a minimum 10 days of information, perhaps more if the 7/20 only includes June.",
      "votes": null
    },
    {
      "id": "1380833",
      "postDate": "07/08/2021 10:55:29",
      "content": "<p>I met the same problem before, perhaps it is because of memory shortage if other have no problems at all.</p>",
      "rawMarkdown": "I met the same problem before, perhaps it is because of memory shortage if other have no problems at all.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1376490,
      "author_name": "julius1r",
      "author_url": "",
      "post_date": "07/05/2021 06:54:16",
      "content": "<p>By the way the link does not work</p>",
      "votes": null,
      "replies": [
        {
          "id": 1377120,
          "author_name": "garggirish",
          "author_url": "",
          "post_date": "07/05/2021 15:37:16",
          "content": "<p>Can you try copy pasting the link in the browser? I think it opens that way</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1377475,
      "author_name": "saimasharleen",
      "author_url": "",
      "post_date": "07/05/2021 22:02:27",
      "content": "<p>You have to keep an eye on your memory usage and maybe it will help you to solve your problem.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1379861,
      "author_name": "andreykrotkikh",
      "author_url": "",
      "post_date": "07/07/2021 16:42:22",
      "content": "<p>Hi, you can pretrain your neural net in the same notebook and save your weights via checkpoints.<br>\n<a href=\"https://www.tensorflow.org/guide/checkpoint\" target=\"_blank\">https://www.tensorflow.org/guide/checkpoint</a></p>\n<p>Then, make next version of your notebook, where you just load weights in the same architecture of net and use pretrained net for submission. So, you will not waste your RAM for training net in submission.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1380321,
          "author_name": "pcjimmmy",
          "author_url": "",
          "post_date": "07/08/2021 03:00:43",
          "content": "<p>A problem with this approach - you will only be able to train your models up to the July 20th update of the training files - so you might miss at a minimum 10 days of information, perhaps more if the 7/20 only includes June.  </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1380833,
      "author_name": "elizabethwang",
      "author_url": "",
      "post_date": "07/08/2021 10:55:29",
      "content": "<p>I met the same problem before, perhaps it is because of memory shortage if other have no problems at all.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1375389": "Hi,\n\n The code seems to run fine on the Kaggle end. After successful commit, If I try to submit it to the competition I get the error \"Notebook Exceeded Allowed Compute\". Can someone help me out figure what is going wrong? \n\nThanks in advance.\n\nMy Notebook link:\n[https://www.kaggle.com/garggirish/mlb-basic-model-1](url)\n\nI've build a very basic model with randomly chosen features. The idea is that first I want to build a very basic model and work my way upwards from there.",
    "1376490": "By the way the link does not work",
    "1377120": "Can you try copy pasting the link in the browser? I think it opens that way",
    "1377475": "You have to keep an eye on your memory usage and maybe it will help you to solve your problem.",
    "1379861": "Hi, you can pretrain your neural net in the same notebook and save your weights via checkpoints.\nhttps://www.tensorflow.org/guide/checkpoint\n\nThen, make next version of your notebook, where you just load weights in the same architecture of net and use pretrained net for submission. So, you will not waste your RAM for training net in submission.",
    "1380321": "A problem with this approach - you will only be able to train your models up to the July 20th update of the training files - so you might miss at a minimum 10 days of information, perhaps more if the 7/20 only includes June.",
    "1380833": "I met the same problem before, perhaps it is because of memory shortage if other have no problems at all."
  },
  "source": "meta"
}