{
  "id": 543875,
  "title": "What is a \"Notebook Inference Server Error\"?",
  "url": "/competitions/jane-street-real-time-market-data-forecasting/discussion/543875",
  "author_name": "",
  "post_date": "2024-11-02T00:03:22.932343700Z",
  "votes": 4,
  "comment_count": 5,
  "views": 0,
  "content": "<p>I am submitting my model but most of the time it gives the weird error: \"Notebook Inference Server Error\". Has anyone had this before?</p>",
  "messages": [
    {
      "id": "3034264",
      "postDate": "11/02/2024 00:03:22",
      "content": "<p>I am submitting my model but most of the time it gives the weird error: \"Notebook Inference Server Error\". Has anyone had this before?</p>",
      "rawMarkdown": "I am submitting my model but most of the time it gives the weird error: \"Notebook Inference Server Error\". Has anyone had this before?",
      "votes": null
    },
    {
      "id": "3035793",
      "postDate": "11/03/2024 20:56:26",
      "content": "<p>I am getting the same error!!</p>",
      "rawMarkdown": "I am getting the same error!!",
      "votes": null
    },
    {
      "id": "3039358",
      "postDate": "11/08/2024 00:34:14",
      "content": "<p>I also get this constantly. Sometimes it tries to run before my packages have been installed.</p>\n<p>I submitted the exact same notebook 5 times yesterday and 4 failed within 60s and then one ran for 9 hours and timed out. I'm not even running a DL inference model either. Literally the most basic linear regression predictor imaginable just to get a submission loading and it still timed out.</p>\n<p>Tearing my hair out trying to understand these submission errors because I can run my own local synthetic test server just fine</p>",
      "rawMarkdown": "I also get this constantly. Sometimes it tries to run before my packages have been installed.\n\nI submitted the exact same notebook 5 times yesterday and 4 failed within 60s and then one ran for 9 hours and timed out. I'm not even running a DL inference model either. Literally the most basic linear regression predictor imaginable just to get a submission loading and it still timed out.\n\nTearing my hair out trying to understand these submission errors because I can run my own local synthetic test server just fine",
      "votes": null
    },
    {
      "id": "3047555",
      "postDate": "11/16/2024 20:40:21",
      "content": "<p>I am facing a similar issue, did you find any solution for that? My submission fails &lt;60s when using pip installs in the package manager and works without them</p>",
      "rawMarkdown": "I am facing a similar issue, did you find any solution for that? My submission fails <60s when using pip installs in the package manager and works without them",
      "votes": null
    },
    {
      "id": "3049344",
      "postDate": "11/19/2024 02:03:10",
      "content": "<p>Yes. I had to save a version of the notebook first. Wait for that to fully build (install packages and run the cells etc). Then when that was completed hit the submission button. For some reason hitting the submission but directly from the edit notebook view tries to run your submission before saving dependencies.</p>\n<p>This is my first time using Kaggle so I'm not really sure why</p>",
      "rawMarkdown": "Yes. I had to save a version of the notebook first. Wait for that to fully build (install packages and run the cells etc). Then when that was completed hit the submission button. For some reason hitting the submission but directly from the edit notebook view tries to run your submission before saving dependencies.\n\nThis is my first time using Kaggle so I'm not really sure why",
      "votes": null
    },
    {
      "id": "3092898",
      "postDate": "01/10/2025 07:45:37",
      "content": "<p>the time between the server streaming its batch to you, and your <code>predict()</code> returns takes more than 900s, so the server times out</p>",
      "rawMarkdown": "the time between the server streaming its batch to you, and your `predict()` returns takes more than 900s, so the server times out",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3035793,
      "author_name": "sankalpbhoyar1402",
      "author_url": "",
      "post_date": "11/03/2024 20:56:26",
      "content": "<p>I am getting the same error!!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3039358,
      "author_name": "michaeltimbs",
      "author_url": "",
      "post_date": "11/08/2024 00:34:14",
      "content": "<p>I also get this constantly. Sometimes it tries to run before my packages have been installed.</p>\n<p>I submitted the exact same notebook 5 times yesterday and 4 failed within 60s and then one ran for 9 hours and timed out. I'm not even running a DL inference model either. Literally the most basic linear regression predictor imaginable just to get a submission loading and it still timed out.</p>\n<p>Tearing my hair out trying to understand these submission errors because I can run my own local synthetic test server just fine</p>",
      "votes": null,
      "replies": [
        {
          "id": 3047555,
          "author_name": "chrthr",
          "author_url": "",
          "post_date": "11/16/2024 20:40:21",
          "content": "<p>I am facing a similar issue, did you find any solution for that? My submission fails &lt;60s when using pip installs in the package manager and works without them</p>",
          "votes": null,
          "replies": [
            {
              "id": 3049344,
              "author_name": "michaeltimbs",
              "author_url": "",
              "post_date": "11/19/2024 02:03:10",
              "content": "<p>Yes. I had to save a version of the notebook first. Wait for that to fully build (install packages and run the cells etc). Then when that was completed hit the submission button. For some reason hitting the submission but directly from the edit notebook view tries to run your submission before saving dependencies.</p>\n<p>This is my first time using Kaggle so I'm not really sure why</p>",
              "votes": null,
              "replies": []
            }
          ]
        },
        {
          "id": 3092898,
          "author_name": "alexmason11",
          "author_url": "",
          "post_date": "01/10/2025 07:45:37",
          "content": "<p>the time between the server streaming its batch to you, and your <code>predict()</code> returns takes more than 900s, so the server times out</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3034264": "I am submitting my model but most of the time it gives the weird error: \"Notebook Inference Server Error\". Has anyone had this before?",
    "3035793": "I am getting the same error!!",
    "3039358": "I also get this constantly. Sometimes it tries to run before my packages have been installed.\n\nI submitted the exact same notebook 5 times yesterday and 4 failed within 60s and then one ran for 9 hours and timed out. I'm not even running a DL inference model either. Literally the most basic linear regression predictor imaginable just to get a submission loading and it still timed out.\n\nTearing my hair out trying to understand these submission errors because I can run my own local synthetic test server just fine",
    "3047555": "I am facing a similar issue, did you find any solution for that? My submission fails <60s when using pip installs in the package manager and works without them",
    "3049344": "Yes. I had to save a version of the notebook first. Wait for that to fully build (install packages and run the cells etc). Then when that was completed hit the submission button. For some reason hitting the submission but directly from the edit notebook view tries to run your submission before saving dependencies.\n\nThis is my first time using Kaggle so I'm not really sure why",
    "3092898": "the time between the server streaming its batch to you, and your `predict()` returns takes more than 900s, so the server times out"
  },
  "source": "meta"
}