{
  "id": 282270,
  "title": "Help for submission issue using Keras model ",
  "url": "/competitions/sartorius-cell-instance-segmentation/discussion/282270",
  "author_name": "",
  "post_date": "2021-10-26T18:39:27.435479200Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hello. I recently tried to submit results using a  Keras UNet model, see <a href=\"https://www.kaggle.com/aramos/sartorius-competition-training-keras-unet\" target=\"_blank\">example notebook</a>. Submission fails, and it is not easy, at least for me, to find and correct the error ('Notebook Threw Exception' /  Error: No additional details provided for this error). </p>\n<p>It seems that the Keras model should incorporate mask overlap correction as mentioned in the description, and then the mask will have the correct format in order to make predictions in the hidden test set.  I am not sure if this is the correct approach or how to implement it inside the model or if there is a  better way to make the submission work. I really appreciate any suggestions or ideas on how to solve this 😊. </p>",
  "messages": [
    {
      "id": "1559138",
      "postDate": "10/26/2021 18:39:27",
      "content": "<p>Hello. I recently tried to submit results using a  Keras UNet model, see <a href=\"https://www.kaggle.com/aramos/sartorius-competition-training-keras-unet\" target=\"_blank\">example notebook</a>. Submission fails, and it is not easy, at least for me, to find and correct the error ('Notebook Threw Exception' /  Error: No additional details provided for this error). </p>\n<p>It seems that the Keras model should incorporate mask overlap correction as mentioned in the description, and then the mask will have the correct format in order to make predictions in the hidden test set.  I am not sure if this is the correct approach or how to implement it inside the model or if there is a  better way to make the submission work. I really appreciate any suggestions or ideas on how to solve this 😊. </p>",
      "rawMarkdown": "Hello. I recently tried to submit results using a  Keras UNet model, see [example notebook](https://www.kaggle.com/aramos/sartorius-competition-training-keras-unet). Submission fails, and it is not easy, at least for me, to find and correct the error ('Notebook Threw Exception' /  Error: No additional details provided for this error). \n\nIt seems that the Keras model should incorporate mask overlap correction as mentioned in the description, and then the mask will have the correct format in order to make predictions in the hidden test set.  I am not sure if this is the correct approach or how to implement it inside the model or if there is a  better way to make the submission work. I really appreciate any suggestions or ideas on how to solve this 😊.",
      "votes": null
    },
    {
      "id": "1559184",
      "postDate": "10/26/2021 19:57:30",
      "content": "<p>Did you create a separate inference kernel with internet off?</p>",
      "rawMarkdown": "Did you create a separate inference kernel with internet off?",
      "votes": null
    },
    {
      "id": "1559200",
      "postDate": "10/26/2021 20:34:41",
      "content": "<p>I simply exported the model as a dataset then loaded it to an inference notebook (with internet off) and it worked . Thanks for your help.  </p>",
      "rawMarkdown": "I simply exported the model as a dataset then loaded it to an inference notebook (with internet off) and it worked . Thanks for your help.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1559184,
      "author_name": "dskswu",
      "author_url": "",
      "post_date": "10/26/2021 19:57:30",
      "content": "<p>Did you create a separate inference kernel with internet off?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1559200,
      "author_name": "aramos",
      "author_url": "",
      "post_date": "10/26/2021 20:34:41",
      "content": "<p>I simply exported the model as a dataset then loaded it to an inference notebook (with internet off) and it worked . Thanks for your help.  </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1559138": "Hello. I recently tried to submit results using a  Keras UNet model, see [example notebook](https://www.kaggle.com/aramos/sartorius-competition-training-keras-unet). Submission fails, and it is not easy, at least for me, to find and correct the error ('Notebook Threw Exception' /  Error: No additional details provided for this error). \n\nIt seems that the Keras model should incorporate mask overlap correction as mentioned in the description, and then the mask will have the correct format in order to make predictions in the hidden test set.  I am not sure if this is the correct approach or how to implement it inside the model or if there is a  better way to make the submission work. I really appreciate any suggestions or ideas on how to solve this 😊.",
    "1559184": "Did you create a separate inference kernel with internet off?",
    "1559200": "I simply exported the model as a dataset then loaded it to an inference notebook (with internet off) and it worked . Thanks for your help."
  },
  "source": "meta"
}