{
  "id": 74953,
  "title": "kernel appears to have died",
  "url": "/competitions/quora-insincere-questions-classification/discussion/74953",
  "author_name": "",
  "post_date": "2018-12-17T14:21:44.116067300Z",
  "votes": null,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Running a pytorch LSTM with a embedding layer, with a vocab of &lt; 8K, dimension of 256 and I receive the error \"The kernel appears to have died.  Restarting kernel automatically.\"  In this particular case, the RAM was well below the allocated 14GB (5.6GB).  Otherwise, I see no other error in the output.</p>\n\n<p>In my working directory i see a core dump equal to max disk size.</p>\n\n<p>This is well below the Glove pretrained weights (vocab &gt; 2M, dim 300) - so I am puzzled as to why this could be so.  </p>\n\n<p>Has anyone had a similar error and can provide some advice in correcting or troubleshooting?</p>",
  "messages": [
    {
      "id": "440413",
      "postDate": "12/17/2018 14:21:44",
      "content": "<p>Running a pytorch LSTM with a embedding layer, with a vocab of &lt; 8K, dimension of 256 and I receive the error \"The kernel appears to have died.  Restarting kernel automatically.\"  In this particular case, the RAM was well below the allocated 14GB (5.6GB).  Otherwise, I see no other error in the output.</p>\n\n<p>In my working directory i see a core dump equal to max disk size.</p>\n\n<p>This is well below the Glove pretrained weights (vocab &gt; 2M, dim 300) - so I am puzzled as to why this could be so.  </p>\n\n<p>Has anyone had a similar error and can provide some advice in correcting or troubleshooting?</p>",
      "rawMarkdown": "Running a pytorch LSTM with a embedding layer, with a vocab of &lt; 8K, dimension of 256 and I receive the error \"The kernel appears to have died.  Restarting kernel automatically.\"  In this particular case, the RAM was well below the allocated 14GB (5.6GB).  Otherwise, I see no other error in the output.\n\nIn my working directory i see a core dump equal to max disk size.\n\nThis is well below the Glove pretrained weights (vocab &gt; 2M, dim 300) - so I am puzzled as to why this could be so.  \n\nHas anyone had a similar error and can provide some advice in correcting or troubleshooting?",
      "votes": null
    },
    {
      "id": "440572",
      "postDate": "12/17/2018 18:08:12",
      "content": "<p>This happens a lot with my setup (Ubuntu Linux 18.10). When it does I type <code>nvidia-smi</code> and kill the nvidia processes and restart the notebook.</p>",
      "rawMarkdown": "This happens a lot with my setup (Ubuntu Linux 18.10). When it does I type `nvidia-smi` and kill the nvidia processes and restart the notebook.",
      "votes": null
    },
    {
      "id": "440642",
      "postDate": "12/17/2018 20:21:11",
      "content": "<p>I can restart the notebook, but it crashes every time.  The identical code, as a python script, runs fine on my Windows 10 laptop which has similar RAM and a much weaker GPU (960M).  I need help in preventing it from happening.</p>",
      "rawMarkdown": "I can restart the notebook, but it crashes every time.  The identical code, as a python script, runs fine on my Windows 10 laptop which has similar RAM and a much weaker GPU (960M).  I need help in preventing it from happening.",
      "votes": null
    },
    {
      "id": "441763",
      "postDate": "12/19/2018 03:05:16",
      "content": "<p>I have meet the similar situation, maybe the code has errors, check carefully.</p>",
      "rawMarkdown": "I have meet the similar situation, maybe the code has errors, check carefully.",
      "votes": null
    },
    {
      "id": "447804",
      "postDate": "12/30/2018 15:44:17",
      "content": "<p>No, it is the exact same code on my Win10 machine.  Does not work in the Kaggle kernel.  How am I supposed to develop anything?  I doubt I will ever compete in a kernel contest again.</p>",
      "rawMarkdown": "No, it is the exact same code on my Win10 machine.  Does not work in the Kaggle kernel.  How am I supposed to develop anything?  I doubt I will ever compete in a kernel contest again.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 440572,
      "author_name": "spirosrap",
      "author_url": "",
      "post_date": "12/17/2018 18:08:12",
      "content": "<p>This happens a lot with my setup (Ubuntu Linux 18.10). When it does I type <code>nvidia-smi</code> and kill the nvidia processes and restart the notebook.</p>",
      "votes": null,
      "replies": [
        {
          "id": 440642,
          "author_name": "arkerpay",
          "author_url": "",
          "post_date": "12/17/2018 20:21:11",
          "content": "<p>I can restart the notebook, but it crashes every time.  The identical code, as a python script, runs fine on my Windows 10 laptop which has similar RAM and a much weaker GPU (960M).  I need help in preventing it from happening.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 441763,
          "author_name": "cutoutsy",
          "author_url": "",
          "post_date": "12/19/2018 03:05:16",
          "content": "<p>I have meet the similar situation, maybe the code has errors, check carefully.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 447804,
          "author_name": "arkerpay",
          "author_url": "",
          "post_date": "12/30/2018 15:44:17",
          "content": "<p>No, it is the exact same code on my Win10 machine.  Does not work in the Kaggle kernel.  How am I supposed to develop anything?  I doubt I will ever compete in a kernel contest again.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "440413": "Running a pytorch LSTM with a embedding layer, with a vocab of &lt; 8K, dimension of 256 and I receive the error \"The kernel appears to have died.  Restarting kernel automatically.\"  In this particular case, the RAM was well below the allocated 14GB (5.6GB).  Otherwise, I see no other error in the output.\n\nIn my working directory i see a core dump equal to max disk size.\n\nThis is well below the Glove pretrained weights (vocab &gt; 2M, dim 300) - so I am puzzled as to why this could be so.  \n\nHas anyone had a similar error and can provide some advice in correcting or troubleshooting?",
    "440572": "This happens a lot with my setup (Ubuntu Linux 18.10). When it does I type `nvidia-smi` and kill the nvidia processes and restart the notebook.",
    "440642": "I can restart the notebook, but it crashes every time.  The identical code, as a python script, runs fine on my Windows 10 laptop which has similar RAM and a much weaker GPU (960M).  I need help in preventing it from happening.",
    "441763": "I have meet the similar situation, maybe the code has errors, check carefully.",
    "447804": "No, it is the exact same code on my Win10 machine.  Does not work in the Kaggle kernel.  How am I supposed to develop anything?  I doubt I will ever compete in a kernel contest again."
  },
  "source": "meta"
}