{
  "id": 133133,
  "title": "Notebook Exceeded Allowed Compute",
  "url": "/competitions/bengaliai-cv19/discussion/133133",
  "author_name": "",
  "post_date": "2020-02-29T22:32:32.277769800Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>This is the general information of my submission kernel....</p>\n\n<ul>\n<li>I'm using an Inference Kernel for submission which just loads model, tests images and generates submission file.</li>\n<li>During the whole execution of notebook, RAM never exceeds 4 gigs</li>\n<li>Whole notebook is executed within 120s</li>\n</ul>\n\n<p>but still I'm facing this(<strong>Notebook Exceeded Allowed Compute</strong>) problem while submission. <strong>What could be the most probable reason?</strong> </p>",
  "messages": [
    {
      "id": "760144",
      "postDate": "02/29/2020 22:32:32",
      "content": "<p>This is the general information of my submission kernel....</p>\n\n<ul>\n<li>I'm using an Inference Kernel for submission which just loads model, tests images and generates submission file.</li>\n<li>During the whole execution of notebook, RAM never exceeds 4 gigs</li>\n<li>Whole notebook is executed within 120s</li>\n</ul>\n\n<p>but still I'm facing this(<strong>Notebook Exceeded Allowed Compute</strong>) problem while submission. <strong>What could be the most probable reason?</strong> </p>",
      "rawMarkdown": "This is the general information of my submission kernel....\n\n- I'm using an Inference Kernel for submission which just loads model, tests images and generates submission file.\n- During the whole execution of notebook, RAM never exceeds 4 gigs\n- Whole notebook is executed within 120s\n\nbut still I'm facing this(**Notebook Exceeded Allowed Compute**) problem while submission. **What could be the most probable reason?**",
      "votes": null
    },
    {
      "id": "760343",
      "postDate": "03/01/2020 06:29:44",
      "content": "<p>During your submission the Kernel is only executed on the example test data - which are like 3 images in total. As such, neither memory nor time will be a problem. During the submission the Kernel will deal with the actual test data, which are about as many images as in the train data. As such, it might cause problems with RAM or time. You can check whether this is the problem by letting your Kernel do inference on the training data, which is roughly the same size.</p>\n\n<p>Apart from this, the error messages seem to be somewhat misleading: Sometimes the problem is actually something else. I believe I got that error myself and it turned out to be an incorrect submission format.</p>",
      "rawMarkdown": "During your submission the Kernel is only executed on the example test data - which are like 3 images in total. As such, neither memory nor time will be a problem. During the submission the Kernel will deal with the actual test data, which are about as many images as in the train data. As such, it might cause problems with RAM or time. You can check whether this is the problem by letting your Kernel do inference on the training data, which is roughly the same size.\n\nApart from this, the error messages seem to be somewhat misleading: Sometimes the problem is actually something else. I believe I got that error myself and it turned out to be an incorrect submission format.",
      "votes": null
    },
    {
      "id": "760401",
      "postDate": "03/01/2020 08:23:06",
      "content": "<p>Thanks. I tested inference kernel on training data and yeah it was getting too heavy for ram.</p>",
      "rawMarkdown": "Thanks. I tested inference kernel on training data and yeah it was getting too heavy for ram.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 760343,
      "author_name": "larswigger",
      "author_url": "",
      "post_date": "03/01/2020 06:29:44",
      "content": "<p>During your submission the Kernel is only executed on the example test data - which are like 3 images in total. As such, neither memory nor time will be a problem. During the submission the Kernel will deal with the actual test data, which are about as many images as in the train data. As such, it might cause problems with RAM or time. You can check whether this is the problem by letting your Kernel do inference on the training data, which is roughly the same size.</p>\n\n<p>Apart from this, the error messages seem to be somewhat misleading: Sometimes the problem is actually something else. I believe I got that error myself and it turned out to be an incorrect submission format.</p>",
      "votes": null,
      "replies": [
        {
          "id": 760401,
          "author_name": "chittalpatel",
          "author_url": "",
          "post_date": "03/01/2020 08:23:06",
          "content": "<p>Thanks. I tested inference kernel on training data and yeah it was getting too heavy for ram.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "760144": "This is the general information of my submission kernel....\n\n- I'm using an Inference Kernel for submission which just loads model, tests images and generates submission file.\n- During the whole execution of notebook, RAM never exceeds 4 gigs\n- Whole notebook is executed within 120s\n\nbut still I'm facing this(**Notebook Exceeded Allowed Compute**) problem while submission. **What could be the most probable reason?**",
    "760343": "During your submission the Kernel is only executed on the example test data - which are like 3 images in total. As such, neither memory nor time will be a problem. During the submission the Kernel will deal with the actual test data, which are about as many images as in the train data. As such, it might cause problems with RAM or time. You can check whether this is the problem by letting your Kernel do inference on the training data, which is roughly the same size.\n\nApart from this, the error messages seem to be somewhat misleading: Sometimes the problem is actually something else. I believe I got that error myself and it turned out to be an incorrect submission format.",
    "760401": "Thanks. I tested inference kernel on training data and yeah it was getting too heavy for ram."
  },
  "source": "meta"
}