{
  "id": 134933,
  "title": "Notebook taking too much time to submit",
  "url": "/competitions/bengaliai-cv19/discussion/134933",
  "author_name": "",
  "post_date": "2020-03-11T07:12:17.314568700Z",
  "votes": null,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi, my notebook is currently taking too long to submit predictions. Any idea what is going wrong?</p>",
  "messages": [
    {
      "id": "768779",
      "postDate": "03/11/2020 07:12:17",
      "content": "<p>Hi, my notebook is currently taking too long to submit predictions. Any idea what is going wrong?</p>",
      "rawMarkdown": "Hi, my notebook is currently taking too long to submit predictions. Any idea what is going wrong?",
      "votes": null
    },
    {
      "id": "768852",
      "postDate": "03/11/2020 09:03:36",
      "content": "<p>Just predict not to train the model. And<a href=\"https://www.kaggle.com/pestipeti/fast-ensemble-5-folds-20-minutes\"> this kernel</a> may help you</p>",
      "rawMarkdown": "Just predict not to train the model. And[ this kernel](https://www.kaggle.com/pestipeti/fast-ensemble-5-folds-20-minutes) may help you",
      "votes": null
    },
    {
      "id": "769007",
      "postDate": "03/11/2020 12:30:56",
      "content": "<p>Same here. Took 4 mins on CPU during commit but 5 hours for the submission....then failed and lost my submissions for today.</p>",
      "rawMarkdown": "Same here. Took 4 mins on CPU during commit but 5 hours for the submission....then failed and lost my submissions for today.",
      "votes": null
    },
    {
      "id": "769010",
      "postDate": "03/11/2020 12:36:42",
      "content": "<p>I think it try to run your kernel once again. and that's the reason why take so much.</p>\n\n<p>So create another kernel and load .h5 models to just predict submission.csv \nand skip training steps in that kernel</p>",
      "rawMarkdown": "I think it try to run your kernel once again. and that's the reason why take so much.\n\nSo create another kernel and load .h5 models to just predict submission.csv \nand skip training steps in that kernel",
      "votes": null
    },
    {
      "id": "769394",
      "postDate": "03/11/2020 21:30:17",
      "content": "<p>When you do the commit the kernel runs only over 12 samples (images), to check that it works at all on very small set.\nIn contrast, when you click \"submit as prediction\" the system runs the kernel over much larger test set (other files are placed instead of small ones).</p>",
      "rawMarkdown": "When you do the commit the kernel runs only over 12 samples (images), to check that it works at all on very small set.\nIn contrast, when you click \"submit as prediction\" the system runs the kernel over much larger test set (other files are placed instead of small ones).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 768852,
      "author_name": "luzerochen",
      "author_url": "",
      "post_date": "03/11/2020 09:03:36",
      "content": "<p>Just predict not to train the model. And<a href=\"https://www.kaggle.com/pestipeti/fast-ensemble-5-folds-20-minutes\"> this kernel</a> may help you</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 769007,
      "author_name": "lpachuong",
      "author_url": "",
      "post_date": "03/11/2020 12:30:56",
      "content": "<p>Same here. Took 4 mins on CPU during commit but 5 hours for the submission....then failed and lost my submissions for today.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 769010,
      "author_name": "masoudmzb",
      "author_url": "",
      "post_date": "03/11/2020 12:36:42",
      "content": "<p>I think it try to run your kernel once again. and that's the reason why take so much.</p>\n\n<p>So create another kernel and load .h5 models to just predict submission.csv \nand skip training steps in that kernel</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 769394,
      "author_name": "dgrechka",
      "author_url": "",
      "post_date": "03/11/2020 21:30:17",
      "content": "<p>When you do the commit the kernel runs only over 12 samples (images), to check that it works at all on very small set.\nIn contrast, when you click \"submit as prediction\" the system runs the kernel over much larger test set (other files are placed instead of small ones).</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "768779": "Hi, my notebook is currently taking too long to submit predictions. Any idea what is going wrong?",
    "768852": "Just predict not to train the model. And[ this kernel](https://www.kaggle.com/pestipeti/fast-ensemble-5-folds-20-minutes) may help you",
    "769007": "Same here. Took 4 mins on CPU during commit but 5 hours for the submission....then failed and lost my submissions for today.",
    "769010": "I think it try to run your kernel once again. and that's the reason why take so much.\n\nSo create another kernel and load .h5 models to just predict submission.csv \nand skip training steps in that kernel",
    "769394": "When you do the commit the kernel runs only over 12 samples (images), to check that it works at all on very small set.\nIn contrast, when you click \"submit as prediction\" the system runs the kernel over much larger test set (other files are placed instead of small ones)."
  },
  "source": "meta"
}