{
  "id": 468436,
  "title": "Submission Score Error",
  "url": "/competitions/blood-vessel-segmentation/discussion/468436",
  "author_name": "",
  "post_date": "2024-01-16T16:43:24.827997700Z",
  "votes": 1,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hello,</p>\n<p>Maybe this would be an inappropriate question, but I still feel I need to ask: If my prediction on the test set aren't 100% matching the expected result, then will I get a Submission Score Error?<br>\nI keep getting this, and I already verified the csv produced, it has exactly the same shape and data type as the one from the sample submission, so I doubt it's that the cause.<br>\nCan anyone please give me at least a hint or tip? Thank you in advance.</p>",
  "messages": [
    {
      "id": "2604788",
      "postDate": "01/16/2024 16:43:24",
      "content": "<p>Hello,</p>\n<p>Maybe this would be an inappropriate question, but I still feel I need to ask: If my prediction on the test set aren't 100% matching the expected result, then will I get a Submission Score Error?<br>\nI keep getting this, and I already verified the csv produced, it has exactly the same shape and data type as the one from the sample submission, so I doubt it's that the cause.<br>\nCan anyone please give me at least a hint or tip? Thank you in advance.</p>",
      "rawMarkdown": "Hello,\n\nMaybe this would be an inappropriate question, but I still feel I need to ask: If my prediction on the test set aren't 100% matching the expected result, then will I get a Submission Score Error?\nI keep getting this, and I already verified the csv produced, it has exactly the same shape and data type as the one from the sample submission, so I doubt it's that the cause.\nCan anyone please give me at least a hint or tip? Thank you in advance.",
      "votes": null
    },
    {
      "id": "2604802",
      "postDate": "01/16/2024 16:54:24",
      "content": "<p>Don't make any assumption about test images rather than they are on /test folder. Make a completely general inference around that.</p>",
      "rawMarkdown": "Don't make any assumption about test images rather than they are on /test folder. Make a completely general inference around that.",
      "votes": null
    },
    {
      "id": "2604954",
      "postDate": "01/16/2024 19:03:35",
      "content": "<p>You might be getting an OOM error if your model is too big and/or you don't resize the test images. </p>",
      "rawMarkdown": "You might be getting an OOM error if your model is too big and/or you don't resize the test images.",
      "votes": null
    },
    {
      "id": "2605718",
      "postDate": "01/17/2024 07:50:24",
      "content": "<p>Thank you for the comment, I already ran the notebook a couple of times (without submitting) and I saw it did well without any issue, but I'll consider rethinking it.</p>",
      "rawMarkdown": "Thank you for the comment, I already ran the notebook a couple of times (without submitting) and I saw it did well without any issue, but I'll consider rethinking it.",
      "votes": null
    },
    {
      "id": "2605719",
      "postDate": "01/17/2024 07:50:43",
      "content": "<p>Ok, thank you for the hint.</p>",
      "rawMarkdown": "Ok, thank you for the hint.",
      "votes": null
    },
    {
      "id": "2605877",
      "postDate": "01/17/2024 09:48:02",
      "content": "<p>Yes, notice that the private test dataset is bigger thus even if it works, it might fail when submitting.</p>",
      "rawMarkdown": "Yes, notice that the private test dataset is bigger thus even if it works, it might fail when submitting.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2604802,
      "author_name": "sacuscreed",
      "author_url": "",
      "post_date": "01/16/2024 16:54:24",
      "content": "<p>Don't make any assumption about test images rather than they are on /test folder. Make a completely general inference around that.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2605719,
          "author_name": "adrian774",
          "author_url": "",
          "post_date": "01/17/2024 07:50:43",
          "content": "<p>Ok, thank you for the hint.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2604954,
      "author_name": "yassinealouini",
      "author_url": "",
      "post_date": "01/16/2024 19:03:35",
      "content": "<p>You might be getting an OOM error if your model is too big and/or you don't resize the test images. </p>",
      "votes": null,
      "replies": [
        {
          "id": 2605718,
          "author_name": "adrian774",
          "author_url": "",
          "post_date": "01/17/2024 07:50:24",
          "content": "<p>Thank you for the comment, I already ran the notebook a couple of times (without submitting) and I saw it did well without any issue, but I'll consider rethinking it.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2605877,
              "author_name": "yassinealouini",
              "author_url": "",
              "post_date": "01/17/2024 09:48:02",
              "content": "<p>Yes, notice that the private test dataset is bigger thus even if it works, it might fail when submitting.</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2604788": "Hello,\n\nMaybe this would be an inappropriate question, but I still feel I need to ask: If my prediction on the test set aren't 100% matching the expected result, then will I get a Submission Score Error?\nI keep getting this, and I already verified the csv produced, it has exactly the same shape and data type as the one from the sample submission, so I doubt it's that the cause.\nCan anyone please give me at least a hint or tip? Thank you in advance.",
    "2604802": "Don't make any assumption about test images rather than they are on /test folder. Make a completely general inference around that.",
    "2604954": "You might be getting an OOM error if your model is too big and/or you don't resize the test images.",
    "2605718": "Thank you for the comment, I already ran the notebook a couple of times (without submitting) and I saw it did well without any issue, but I'll consider rethinking it.",
    "2605719": "Ok, thank you for the hint.",
    "2605877": "Yes, notice that the private test dataset is bigger thus even if it works, it might fail when submitting."
  },
  "source": "meta"
}