{
  "id": 399652,
  "title": "Notebook Out of Memory",
  "url": "/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/399652",
  "author_name": "",
  "post_date": "2023-04-05T01:27:44.991631800Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I submitted my notebook but got this error…..</p>\n<p>It ran fine locally but got \"out of memory\" error after submission….</p>\n<p>I am wondering what is the memory limit for submission? Is it possible to increase the memory limit?</p>",
  "messages": [
    {
      "id": "2209854",
      "postDate": "04/05/2023 01:27:44",
      "content": "<p>I submitted my notebook but got this error…..</p>\n<p>It ran fine locally but got \"out of memory\" error after submission….</p>\n<p>I am wondering what is the memory limit for submission? Is it possible to increase the memory limit?</p>",
      "rawMarkdown": "I submitted my notebook but got this error.....\n\n\nIt ran fine locally but got \"out of memory\" error after submission....\n\nI am wondering what is the memory limit for submission? Is it possible to increase the memory limit?",
      "votes": null
    },
    {
      "id": "2210689",
      "postDate": "04/05/2023 14:59:26",
      "content": "<p>Kaggle Notebooks are given 30 GB RAM for CPU-only notebooks and ~16GB for GPU and TPU notebooks. For interactive notebooks you can view the allocations by clicking the metrics in the top right hand corner and expanding them.</p>\n<p>For one of my notebooks (GPU), I was facing memory issues during submission. Turns out the problem was that my prediction step created a single large Numpy array which was much larger with the hidden data added to test during submission.</p>\n<p>A good test to do for finding the error is to see if trying to do the same prediction step for all train folders (defog, tdscfog, notype) leads to a \"out of memory\" error.</p>",
      "rawMarkdown": "Kaggle Notebooks are given 30 GB RAM for CPU-only notebooks and ~16GB for GPU and TPU notebooks. For interactive notebooks you can view the allocations by clicking the metrics in the top right hand corner and expanding them.\n\nFor one of my notebooks (GPU), I was facing memory issues during submission. Turns out the problem was that my prediction step created a single large Numpy array which was much larger with the hidden data added to test during submission.\n\nA good test to do for finding the error is to see if trying to do the same prediction step for all train folders (defog, tdscfog, notype) leads to a \"out of memory\" error.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2210689,
      "author_name": "coderrkj",
      "author_url": "",
      "post_date": "04/05/2023 14:59:26",
      "content": "<p>Kaggle Notebooks are given 30 GB RAM for CPU-only notebooks and ~16GB for GPU and TPU notebooks. For interactive notebooks you can view the allocations by clicking the metrics in the top right hand corner and expanding them.</p>\n<p>For one of my notebooks (GPU), I was facing memory issues during submission. Turns out the problem was that my prediction step created a single large Numpy array which was much larger with the hidden data added to test during submission.</p>\n<p>A good test to do for finding the error is to see if trying to do the same prediction step for all train folders (defog, tdscfog, notype) leads to a \"out of memory\" error.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2209854": "I submitted my notebook but got this error.....\n\n\nIt ran fine locally but got \"out of memory\" error after submission....\n\nI am wondering what is the memory limit for submission? Is it possible to increase the memory limit?",
    "2210689": "Kaggle Notebooks are given 30 GB RAM for CPU-only notebooks and ~16GB for GPU and TPU notebooks. For interactive notebooks you can view the allocations by clicking the metrics in the top right hand corner and expanding them.\n\nFor one of my notebooks (GPU), I was facing memory issues during submission. Turns out the problem was that my prediction step created a single large Numpy array which was much larger with the hidden data added to test during submission.\n\nA good test to do for finding the error is to see if trying to do the same prediction step for all train folders (defog, tdscfog, notype) leads to a \"out of memory\" error."
  },
  "source": "meta"
}