{
  "id": 104813,
  "title": "Public score 0.000 again and again...",
  "url": "/competitions/aptos2019-blindness-detection/discussion/104813",
  "author_name": "",
  "post_date": "2019-08-19T10:19:54.572730300Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Dear community,</p>\n\n<p>I hope you can help me. Although I managed to submit my csv file to the competition, I've tried everything to get a public score &gt; 0.000 but it's been impossible so far. Can anyone check my <a href=\"https://www.kaggle.com/hefestion77/aptos-final-submission/notebook\">notebook </a> and let me know why it returns a public score of 0.000???</p>\n\n<p>I've tried to substitute the sample_submission.diagnosis column with the predictions gathered as a list of strings, of integers, and also as a numpy array of integers, but I always get a public score of 0.000 whatever the method I use to include my predictions in the diagnosis column...</p>\n\n<p>Thanks in advance!</p>\n\n<p>Sergio.</p>",
  "messages": [
    {
      "id": "602651",
      "postDate": "08/19/2019 10:19:54",
      "content": "<p>Dear community,</p>\n\n<p>I hope you can help me. Although I managed to submit my csv file to the competition, I've tried everything to get a public score &gt; 0.000 but it's been impossible so far. Can anyone check my <a href=\"https://www.kaggle.com/hefestion77/aptos-final-submission/notebook\">notebook </a> and let me know why it returns a public score of 0.000???</p>\n\n<p>I've tried to substitute the sample_submission.diagnosis column with the predictions gathered as a list of strings, of integers, and also as a numpy array of integers, but I always get a public score of 0.000 whatever the method I use to include my predictions in the diagnosis column...</p>\n\n<p>Thanks in advance!</p>\n\n<p>Sergio.</p>",
      "rawMarkdown": "Dear community,\n\nI hope you can help me. Although I managed to submit my csv file to the competition, I've tried everything to get a public score &gt; 0.000 but it's been impossible so far. Can anyone check my [notebook ](https://www.kaggle.com/hefestion77/aptos-final-submission/notebook) and let me know why it returns a public score of 0.000???\n\nI've tried to substitute the sample_submission.diagnosis column with the predictions gathered as a list of strings, of integers, and also as a numpy array of integers, but I always get a public score of 0.000 whatever the method I use to include my predictions in the diagnosis column...\n\nThanks in advance!\n\nSergio.",
      "votes": null
    },
    {
      "id": "605567",
      "postDate": "08/22/2019 13:58:39",
      "content": "<p>I am having the same problem. how do you fix this?</p>",
      "rawMarkdown": "I am having the same problem. how do you fix this?",
      "votes": null
    },
    {
      "id": "606623",
      "postDate": "08/23/2019 20:00:58",
      "content": "<p>Hi,</p>\n\n<p>Checking your link returns a 404. Just trying to submit a valid score took up a lot of my time in this competition. See if you can share the code and maybe we can help. </p>",
      "rawMarkdown": "Hi,\n\nChecking your link returns a 404. Just trying to submit a valid score took up a lot of my time in this competition. See if you can share the code and maybe we can help.",
      "votes": null
    },
    {
      "id": "607241",
      "postDate": "08/24/2019 21:44:01",
      "content": "<p>I fixed it by reading, resizing and predicting for every single image from the test directory in the same notebook which creates the submission file. I don't know why this is the only way to make it work properly.</p>",
      "rawMarkdown": "I fixed it by reading, resizing and predicting for every single image from the test directory in the same notebook which creates the submission file. I don't know why this is the only way to make it work properly.",
      "votes": null
    },
    {
      "id": "607303",
      "postDate": "08/25/2019 01:46:57",
      "content": "<p>You may know this, but when the submission is made, the private test set is substituted in for the public test set, so any code that was running on the public test set must also run on the private. This has the benefit of allowing us to make sure our code is running on the private test set before the competition ends, but it does require the tedious step of debugging on inputs we can't see. </p>",
      "rawMarkdown": "You may know this, but when the submission is made, the private test set is substituted in for the public test set, so any code that was running on the public test set must also run on the private. This has the benefit of allowing us to make sure our code is running on the private test set before the competition ends, but it does require the tedious step of debugging on inputs we can't see.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 605567,
      "author_name": "neeraj17",
      "author_url": "",
      "post_date": "08/22/2019 13:58:39",
      "content": "<p>I am having the same problem. how do you fix this?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 606623,
      "author_name": "amqdnguyen",
      "author_url": "",
      "post_date": "08/23/2019 20:00:58",
      "content": "<p>Hi,</p>\n\n<p>Checking your link returns a 404. Just trying to submit a valid score took up a lot of my time in this competition. See if you can share the code and maybe we can help. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 607241,
      "author_name": "hefestion77",
      "author_url": "",
      "post_date": "08/24/2019 21:44:01",
      "content": "<p>I fixed it by reading, resizing and predicting for every single image from the test directory in the same notebook which creates the submission file. I don't know why this is the only way to make it work properly.</p>",
      "votes": null,
      "replies": [
        {
          "id": 607303,
          "author_name": "amqdnguyen",
          "author_url": "",
          "post_date": "08/25/2019 01:46:57",
          "content": "<p>You may know this, but when the submission is made, the private test set is substituted in for the public test set, so any code that was running on the public test set must also run on the private. This has the benefit of allowing us to make sure our code is running on the private test set before the competition ends, but it does require the tedious step of debugging on inputs we can't see. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "602651": "Dear community,\n\nI hope you can help me. Although I managed to submit my csv file to the competition, I've tried everything to get a public score &gt; 0.000 but it's been impossible so far. Can anyone check my [notebook ](https://www.kaggle.com/hefestion77/aptos-final-submission/notebook) and let me know why it returns a public score of 0.000???\n\nI've tried to substitute the sample_submission.diagnosis column with the predictions gathered as a list of strings, of integers, and also as a numpy array of integers, but I always get a public score of 0.000 whatever the method I use to include my predictions in the diagnosis column...\n\nThanks in advance!\n\nSergio.",
    "605567": "I am having the same problem. how do you fix this?",
    "606623": "Hi,\n\nChecking your link returns a 404. Just trying to submit a valid score took up a lot of my time in this competition. See if you can share the code and maybe we can help.",
    "607241": "I fixed it by reading, resizing and predicting for every single image from the test directory in the same notebook which creates the submission file. I don't know why this is the only way to make it work properly.",
    "607303": "You may know this, but when the submission is made, the private test set is substituted in for the public test set, so any code that was running on the public test set must also run on the private. This has the benefit of allowing us to make sure our code is running on the private test set before the competition ends, but it does require the tedious step of debugging on inputs we can't see."
  },
  "source": "meta"
}