{
  "id": 125689,
  "title": "Problems with kernel commiting",
  "url": "/competitions/bengaliai-cv19/discussion/125689",
  "author_name": "",
  "post_date": "2020-01-12T21:38:50.021331900Z",
  "votes": 1,
  "comment_count": 9,
  "views": 0,
  "content": "<p>It's strange: my kernel works fine when I execute all code cell by cell but it always fails when I'm trying to commit it.\nCan somebody from Kaggle Team help me with this issue?\nBest regards.</p>",
  "messages": [
    {
      "id": "717208",
      "postDate": "01/12/2020 21:38:50",
      "content": "<p>It's strange: my kernel works fine when I execute all code cell by cell but it always fails when I'm trying to commit it.\nCan somebody from Kaggle Team help me with this issue?\nBest regards.</p>",
      "rawMarkdown": "It's strange: my kernel works fine when I execute all code cell by cell but it always fails when I'm trying to commit it.\nCan somebody from Kaggle Team help me with this issue?\nBest regards.",
      "votes": null
    },
    {
      "id": "717251",
      "postDate": "01/12/2020 23:32:55",
      "content": "<p>if you create local folder or files make sure to delete them at the end of notebook </p>",
      "rawMarkdown": "if you create local folder or files make sure to delete them at the end of notebook",
      "votes": null
    },
    {
      "id": "717924",
      "postDate": "01/13/2020 19:34:39",
      "content": "<p>Thank you for suggestion.</p>",
      "rawMarkdown": "Thank you for suggestion.",
      "votes": null
    },
    {
      "id": "719824",
      "postDate": "01/15/2020 22:12:57",
      "content": "<p>I didn't help me either...</p>",
      "rawMarkdown": "I didn't help me either...",
      "votes": null
    },
    {
      "id": "721593",
      "postDate": "01/17/2020 14:15:47",
      "content": "<p>Found out that this error happens whe I'm fitting model:\n<strong>Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled</strong>\nDoes anybody know how to cure it?</p>",
      "rawMarkdown": "Found out that this error happens whe I'm fitting model:\n**Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled**\nDoes anybody know how to cure it?",
      "votes": null
    },
    {
      "id": "722919",
      "postDate": "01/19/2020 09:58:19",
      "content": "<p>Looks like model.fit method in tensorflow is caching training dataset in ram and it fails kernel in the end even if you use gpu.\nUnfortunatelly, it's still not solved.\n<a href=\"https://github.com/tensorflow/tensorflow/issues/31312\">https://github.com/tensorflow/tensorflow/issues/31312</a></p>",
      "rawMarkdown": "Looks like model.fit method in tensorflow is caching training dataset in ram and it fails kernel in the end even if you use gpu.\nUnfortunatelly, it's still not solved.\nhttps://github.com/tensorflow/tensorflow/issues/31312",
      "votes": null
    },
    {
      "id": "727566",
      "postDate": "01/23/2020 20:38:05",
      "content": "<p>Solved by directly import ImageDataGenerator from tensorflow.keras.preprocessing.images instead of using full path for class rewriting.</p>",
      "rawMarkdown": "Solved by directly import ImageDataGenerator from tensorflow.keras.preprocessing.images instead of using full path for class rewriting.",
      "votes": null
    },
    {
      "id": "729069",
      "postDate": "01/25/2020 17:33:45",
      "content": "<p>So why didn't it cause the error when running the code before committing? </p>",
      "rawMarkdown": "So why didn't it cause the error when running the code before committing?",
      "votes": null
    },
    {
      "id": "736500",
      "postDate": "02/04/2020 09:22:00",
      "content": "<p>Don't know. Nobody helped me so I've been deleting and running row by row to get that exception.</p>",
      "rawMarkdown": "Don't know. Nobody helped me so I've been deleting and running row by row to get that exception.",
      "votes": null
    },
    {
      "id": "737817",
      "postDate": "02/05/2020 20:01:02",
      "content": "<p>Solved by:</p>\n\n<p>import tensorflow as tf\ntf.keras.backend.clear_session()</p>",
      "rawMarkdown": "Solved by:\n\nimport tensorflow as tf\ntf.keras.backend.clear_session()",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 717251,
      "author_name": "drhabib",
      "author_url": "",
      "post_date": "01/12/2020 23:32:55",
      "content": "<p>if you create local folder or files make sure to delete them at the end of notebook </p>",
      "votes": null,
      "replies": [
        {
          "id": 717924,
          "author_name": "dishkakrauch",
          "author_url": "",
          "post_date": "01/13/2020 19:34:39",
          "content": "<p>Thank you for suggestion.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 719824,
          "author_name": "dishkakrauch",
          "author_url": "",
          "post_date": "01/15/2020 22:12:57",
          "content": "<p>I didn't help me either...</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 721593,
      "author_name": "dishkakrauch",
      "author_url": "",
      "post_date": "01/17/2020 14:15:47",
      "content": "<p>Found out that this error happens whe I'm fitting model:\n<strong>Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled</strong>\nDoes anybody know how to cure it?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 722919,
      "author_name": "dishkakrauch",
      "author_url": "",
      "post_date": "01/19/2020 09:58:19",
      "content": "<p>Looks like model.fit method in tensorflow is caching training dataset in ram and it fails kernel in the end even if you use gpu.\nUnfortunatelly, it's still not solved.\n<a href=\"https://github.com/tensorflow/tensorflow/issues/31312\">https://github.com/tensorflow/tensorflow/issues/31312</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 727566,
      "author_name": "dishkakrauch",
      "author_url": "",
      "post_date": "01/23/2020 20:38:05",
      "content": "<p>Solved by directly import ImageDataGenerator from tensorflow.keras.preprocessing.images instead of using full path for class rewriting.</p>",
      "votes": null,
      "replies": [
        {
          "id": 729069,
          "author_name": "kokecacao",
          "author_url": "",
          "post_date": "01/25/2020 17:33:45",
          "content": "<p>So why didn't it cause the error when running the code before committing? </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 736500,
          "author_name": "dishkakrauch",
          "author_url": "",
          "post_date": "02/04/2020 09:22:00",
          "content": "<p>Don't know. Nobody helped me so I've been deleting and running row by row to get that exception.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 737817,
      "author_name": "dishkakrauch",
      "author_url": "",
      "post_date": "02/05/2020 20:01:02",
      "content": "<p>Solved by:</p>\n\n<p>import tensorflow as tf\ntf.keras.backend.clear_session()</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "717208": "It's strange: my kernel works fine when I execute all code cell by cell but it always fails when I'm trying to commit it.\nCan somebody from Kaggle Team help me with this issue?\nBest regards.",
    "717251": "if you create local folder or files make sure to delete them at the end of notebook",
    "717924": "Thank you for suggestion.",
    "719824": "I didn't help me either...",
    "721593": "Found out that this error happens whe I'm fitting model:\n**Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled**\nDoes anybody know how to cure it?",
    "722919": "Looks like model.fit method in tensorflow is caching training dataset in ram and it fails kernel in the end even if you use gpu.\nUnfortunatelly, it's still not solved.\nhttps://github.com/tensorflow/tensorflow/issues/31312",
    "727566": "Solved by directly import ImageDataGenerator from tensorflow.keras.preprocessing.images instead of using full path for class rewriting.",
    "729069": "So why didn't it cause the error when running the code before committing?",
    "736500": "Don't know. Nobody helped me so I've been deleting and running row by row to get that exception.",
    "737817": "Solved by:\n\nimport tensorflow as tf\ntf.keras.backend.clear_session()"
  },
  "source": "meta"
}