{
  "id": 532417,
  "title": "Submission Scoring Error",
  "url": "/competitions/ariel-data-challenge-2024/discussion/532417",
  "author_name": "Artem Shlezinger",
  "post_date": "2024-09-06T08:53:58.486000",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>After waiting over one hour (actually twice), I got an error \"submission scoring error,\" which, according to the Kaggle documentation, <br>\n\"Your notebook generated a submission file with incorrect format. Some examples causing this are: the wrong number of rows or columns, empty values, an incorrect data type for a value, or invalid submission values from what is expected.\" </p>\n<p>I've double-checked my submission.csv N times and it is completely identical to the sample_submission.csv structure.</p>\n<table>\n<thead>\n<tr>\n<th>planet_id</th>\n<th>wl_1</th>\n<th>wl_2</th>\n<th>wl_3</th>\n<th>wl_4</th>\n<th>wl_5</th>\n<th>…</th>\n<th>sigma_283</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>499191466</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>…</td>\n<td>0.00327921930313997</td>\n</tr>\n</tbody>\n</table>\n<p>To form csv structure, I'm using:<br>\n<code>cols = pd.read_csv(f\"{ROOT}/sample_submission.csv\").columns</code></p>\n<p>The shape of the result dataset is (n, 567) (1 column = player_id; 283 cols = spectra; 283 cols = uncertainty)</p>\n<p>Has someone encountered this (or similar) issue when submitting the notebook?</p>\n<p>I've also hardcoded σ to avoid my model generating a 0-value (which could lead to division by 0 using the evaluating formula) but it didn't help either. </p>",
  "messages": [
    {
      "id": 2980926,
      "postDate": "2024-09-06T10:07:18.717Z",
      "content": "<p>Try clipping negative value to zero.</p>",
      "rawMarkdown": "Try clipping negative value to zero.",
      "votes": 1
    },
    {
      "id": 2980917,
      "postDate": "2024-09-06T09:51:02.110Z",
      "content": "<p>Try to insert some fake test records to check the shapes.Some problems are not evident if there is only 1 test rec.(At least in my case)<br>\nCheck memory at inference. Kaggle masks all errors as 'submission scoring errors' , which is very misleading</p>",
      "rawMarkdown": "Try to insert some fake test records to check the shapes.Some problems are not evident if there is only 1 test rec.(At least in my case)\nCheck memory at inference. Kaggle masks all errors as 'submission scoring errors' , which is very misleading",
      "votes": 1
    },
    {
      "id": 2980879,
      "postDate": "2024-09-06T08:53:58.487Z",
      "content": "<p>After waiting over one hour (actually twice), I got an error \"submission scoring error,\" which, according to the Kaggle documentation, <br>\n\"Your notebook generated a submission file with incorrect format. Some examples causing this are: the wrong number of rows or columns, empty values, an incorrect data type for a value, or invalid submission values from what is expected.\" </p>\n<p>I've double-checked my submission.csv N times and it is completely identical to the sample_submission.csv structure.</p>\n<table>\n<thead>\n<tr>\n<th>planet_id</th>\n<th>wl_1</th>\n<th>wl_2</th>\n<th>wl_3</th>\n<th>wl_4</th>\n<th>wl_5</th>\n<th>…</th>\n<th>sigma_283</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>499191466</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>0.00327921930313997</td>\n<td>…</td>\n<td>0.00327921930313997</td>\n</tr>\n</tbody>\n</table>\n<p>To form csv structure, I'm using:<br>\n<code>cols = pd.read_csv(f\"{ROOT}/sample_submission.csv\").columns</code></p>\n<p>The shape of the result dataset is (n, 567) (1 column = player_id; 283 cols = spectra; 283 cols = uncertainty)</p>\n<p>Has someone encountered this (or similar) issue when submitting the notebook?</p>\n<p>I've also hardcoded σ to avoid my model generating a 0-value (which could lead to division by 0 using the evaluating formula) but it didn't help either. </p>",
      "rawMarkdown": "After waiting over one hour (actually twice), I got an error \"submission scoring error,\" which, according to the Kaggle documentation, \n\"Your notebook generated a submission file with incorrect format. Some examples causing this are: the wrong number of rows or columns, empty values, an incorrect data type for a value, or invalid submission values from what is expected.\" \n\nI've double-checked my submission.csv N times and it is completely identical to the sample_submission.csv structure.\n\n| planet_id|wl_1|wl_2|wl_3|wl_4|wl_5|...|sigma_283|\n| --- | --- | --- | --- | --- | --- | --- | --- |\n499191466|0.00327921930313997|0.00327921930313997|0.00327921930313997|0.00327921930313997|0.00327921930313997|...|0.00327921930313997|\n\nTo form csv structure, I'm using:\n`cols = pd.read_csv(f\"{ROOT}/sample_submission.csv\").columns`\n\nThe shape of the result dataset is (n, 567) (1 column = player_id; 283 cols = spectra; 283 cols = uncertainty)\n\nHas someone encountered this (or similar) issue when submitting the notebook?\n\nI've also hardcoded σ to avoid my model generating a 0-value (which could lead to division by 0 using the evaluating formula) but it didn't help either. ",
      "votes": 1
    },
    {
      "id": 2980948,
      "postDate": "2024-09-06T10:27:39.803Z",
      "content": "<p>Thank you very much for the responses 🙇‍♂️<br>\nThe steps undertaken so far:</p>\n<ul>\n<li>composed submission.csv with completely hardcoded values -&gt; successfully submitted with the score 0 🙂</li>\n<li>added handling of extreme values (0, negative values, bigger than a specified threshold etc.) generated by a model -&gt; since I've already used 3 attempts this fix will be evaluated tomorrow 🙂</li>\n</ul>\n<p>I can infer that the structure of my csv file is correct (relying on p.1) and the reason obviously lies in my weak model (I just wanted to submit anything to go through the full process using a basic ML algorithm)</p>\n<p>*This is very inconvenient to wait 1 hour to get an error without any details and in addition get a reduction in the number of attempts🙁</p>",
      "rawMarkdown": "Thank you very much for the responses 🙇‍♂️\nThe steps undertaken so far:\n- composed submission.csv with completely hardcoded values -> successfully submitted with the score 0 🙂\n- added handling of extreme values (0, negative values, bigger than a specified threshold etc.) generated by a model -> since I've already used 3 attempts this fix will be evaluated tomorrow 🙂\n\nI can infer that the structure of my csv file is correct (relying on p.1) and the reason obviously lies in my weak model (I just wanted to submit anything to go through the full process using a basic ML algorithm)\n\n\n*This is very inconvenient to wait 1 hour to get an error without any details and in addition get a reduction in the number of attempts🙁\n"
    }
  ],
  "comments": [
    {
      "id": 2980926,
      "author_name": "ChingYinNg",
      "author_url": "",
      "post_date": "2024-09-06T10:07:18.717000",
      "content": "<p>Try clipping negative value to zero.</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 2980917,
      "author_name": "Evdilos_Ikaria",
      "author_url": "",
      "post_date": "2024-09-06T09:51:02.110000",
      "content": "<p>Try to insert some fake test records to check the shapes.Some problems are not evident if there is only 1 test rec.(At least in my case)<br>\nCheck memory at inference. Kaggle masks all errors as 'submission scoring errors' , which is very misleading</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 2980948,
      "author_name": "Artem Shlezinger",
      "author_url": "",
      "post_date": "2024-09-06T10:27:39.803000",
      "content": "<p>Thank you very much for the responses 🙇‍♂️<br>\nThe steps undertaken so far:</p>\n<ul>\n<li>composed submission.csv with completely hardcoded values -&gt; successfully submitted with the score 0 🙂</li>\n<li>added handling of extreme values (0, negative values, bigger than a specified threshold etc.) generated by a model -&gt; since I've already used 3 attempts this fix will be evaluated tomorrow 🙂</li>\n</ul>\n<p>I can infer that the structure of my csv file is correct (relying on p.1) and the reason obviously lies in my weak model (I just wanted to submit anything to go through the full process using a basic ML algorithm)</p>\n<p>*This is very inconvenient to wait 1 hour to get an error without any details and in addition get a reduction in the number of attempts🙁</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2980926": "Try clipping negative value to zero.",
    "2980917": "Try to insert some fake test records to check the shapes.Some problems are not evident if there is only 1 test rec.(At least in my case)\nCheck memory at inference. Kaggle masks all errors as 'submission scoring errors' , which is very misleading",
    "2980879": "After waiting over one hour (actually twice), I got an error \"submission scoring error,\" which, according to the Kaggle documentation, \n\"Your notebook generated a submission file with incorrect format. Some examples causing this are: the wrong number of rows or columns, empty values, an incorrect data type for a value, or invalid submission values from what is expected.\" \n\nI've double-checked my submission.csv N times and it is completely identical to the sample_submission.csv structure.\n\n| planet_id|wl_1|wl_2|wl_3|wl_4|wl_5|...|sigma_283|\n| --- | --- | --- | --- | --- | --- | --- | --- |\n499191466|0.00327921930313997|0.00327921930313997|0.00327921930313997|0.00327921930313997|0.00327921930313997|...|0.00327921930313997|\n\nTo form csv structure, I'm using:\n`cols = pd.read_csv(f\"{ROOT}/sample_submission.csv\").columns`\n\nThe shape of the result dataset is (n, 567) (1 column = player_id; 283 cols = spectra; 283 cols = uncertainty)\n\nHas someone encountered this (or similar) issue when submitting the notebook?\n\nI've also hardcoded σ to avoid my model generating a 0-value (which could lead to division by 0 using the evaluating formula) but it didn't help either. ",
    "2980948": "Thank you very much for the responses 🙇‍♂️\nThe steps undertaken so far:\n- composed submission.csv with completely hardcoded values -> successfully submitted with the score 0 🙂\n- added handling of extreme values (0, negative values, bigger than a specified threshold etc.) generated by a model -> since I've already used 3 attempts this fix will be evaluated tomorrow 🙂\n\nI can infer that the structure of my csv file is correct (relying on p.1) and the reason obviously lies in my weak model (I just wanted to submit anything to go through the full process using a basic ML algorithm)\n\n\n*This is very inconvenient to wait 1 hour to get an error without any details and in addition get a reduction in the number of attempts🙁\n"
  }
}