{
  "id": 77839,
  "title": "What is the reason kernel sometimes dies while loading data?",
  "url": "/competitions/LANL-Earthquake-Prediction/discussion/77839",
  "author_name": "",
  "post_date": "2019-01-17T03:17:13.251539200Z",
  "votes": 1,
  "comment_count": 5,
  "views": 0,
  "content": "<p>I thought kernels have always the same resources, but sometimes I can load data and sometimes kernel dies.</p>\n\n<p>I use the code everyone uses:</p>\n\n<pre><code>train = pd.read_csv('../input/train.csv', dtype={'acoustic_data': np.int16, 'time_to_failure': np.float64})\n</code></pre>\n\n<p>Sometimes it works, sometimes not. Anyone experiences similar problems?</p>",
  "messages": [
    {
      "id": "457164",
      "postDate": "01/17/2019 03:17:13",
      "content": "<p>I thought kernels have always the same resources, but sometimes I can load data and sometimes kernel dies.</p>\n\n<p>I use the code everyone uses:</p>\n\n<pre><code>train = pd.read_csv('../input/train.csv', dtype={'acoustic_data': np.int16, 'time_to_failure': np.float64})\n</code></pre>\n\n<p>Sometimes it works, sometimes not. Anyone experiences similar problems?</p>",
      "rawMarkdown": "I thought kernels have always the same resources, but sometimes I can load data and sometimes kernel dies.\n\nI use the code everyone uses:\n\n    train = pd.read_csv('../input/train.csv', dtype={'acoustic_data': np.int16, 'time_to_failure': np.float64})\n\nSometimes it works, sometimes not. Anyone experiences similar problems?",
      "votes": null
    },
    {
      "id": "457215",
      "postDate": "01/17/2019 04:57:28",
      "content": "<p>Your kernel is out of memory, set np.float64 to np.float32</p>",
      "rawMarkdown": "Your kernel is out of memory, set np.float64 to np.float32",
      "votes": null
    },
    {
      "id": "457852",
      "postDate": "01/18/2019 07:56:15",
      "content": "<p>Yes this has happened to me.  Did you find a solution?</p>",
      "rawMarkdown": "Yes this has happened to me.  Did you find a solution?",
      "votes": null
    },
    {
      "id": "457870",
      "postDate": "01/18/2019 08:37:10",
      "content": "<p>Thank you, this helped!</p>",
      "rawMarkdown": "Thank you, this helped!",
      "votes": null
    },
    {
      "id": "457873",
      "postDate": "01/18/2019 08:37:48",
      "content": "<p>yes! As <a href=\"/wimwim\">@wimwim</a> said you need to set np.float32 while loading data</p>",
      "rawMarkdown": "yes! As @wimwim said you need to set np.float32 while loading data",
      "votes": null
    },
    {
      "id": "458127",
      "postDate": "01/18/2019 20:47:50",
      "content": "<p>I have similar problem like you, After every loading data I reload kernel and work from first.</p>",
      "rawMarkdown": "I have similar problem like you, After every loading data I reload kernel and work from first.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 457215,
      "author_name": "wimwim",
      "author_url": "",
      "post_date": "01/17/2019 04:57:28",
      "content": "<p>Your kernel is out of memory, set np.float64 to np.float32</p>",
      "votes": null,
      "replies": [
        {
          "id": 457870,
          "author_name": "artgor",
          "author_url": "",
          "post_date": "01/18/2019 08:37:10",
          "content": "<p>Thank you, this helped!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 457852,
      "author_name": "chceight",
      "author_url": "",
      "post_date": "01/18/2019 07:56:15",
      "content": "<p>Yes this has happened to me.  Did you find a solution?</p>",
      "votes": null,
      "replies": [
        {
          "id": 457873,
          "author_name": "artgor",
          "author_url": "",
          "post_date": "01/18/2019 08:37:48",
          "content": "<p>yes! As <a href=\"/wimwim\">@wimwim</a> said you need to set np.float32 while loading data</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 458127,
      "author_name": "msahebi",
      "author_url": "",
      "post_date": "01/18/2019 20:47:50",
      "content": "<p>I have similar problem like you, After every loading data I reload kernel and work from first.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "457164": "I thought kernels have always the same resources, but sometimes I can load data and sometimes kernel dies.\n\nI use the code everyone uses:\n\n    train = pd.read_csv('../input/train.csv', dtype={'acoustic_data': np.int16, 'time_to_failure': np.float64})\n\nSometimes it works, sometimes not. Anyone experiences similar problems?",
    "457215": "Your kernel is out of memory, set np.float64 to np.float32",
    "457852": "Yes this has happened to me.  Did you find a solution?",
    "457870": "Thank you, this helped!",
    "457873": "yes! As @wimwim said you need to set np.float32 while loading data",
    "458127": "I have similar problem like you, After every loading data I reload kernel and work from first."
  },
  "source": "meta"
}