{
  "id": 191897,
  "title": "Advancing from Example submissions to Actual (private hidden) submissions",
  "url": "/competitions/riiid-test-answer-prediction/discussion/191897",
  "author_name": "",
  "post_date": "2020-10-19T09:40:39.298239Z",
  "votes": null,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Thanks, folks - you advice has been really helpful. Now I want to move on to the real thing ….</p>\n<p>This seems to be the standard instructions:</p>\n<p>iter_test = env.iter_test() <br>\n(loop start)<br>\n(test_df, sample_prediction_df) = next(iter_test)<br>\n….<br>\nenv.predict(my_prediction_for_test_df)<br>\n(loop end)</p>\n<p>How can we tell that test_df is a private/hidden test and not an example test?<br>\nAnd how do we get feedback about the prediction file? I have been submitting Example predictions, but have seen no feedback, such as file-format errors.</p>",
  "messages": [
    {
      "id": "1053732",
      "postDate": "10/19/2020 09:40:39",
      "content": "<p>Thanks, folks - you advice has been really helpful. Now I want to move on to the real thing ….</p>\n<p>This seems to be the standard instructions:</p>\n<p>iter_test = env.iter_test() <br>\n(loop start)<br>\n(test_df, sample_prediction_df) = next(iter_test)<br>\n….<br>\nenv.predict(my_prediction_for_test_df)<br>\n(loop end)</p>\n<p>How can we tell that test_df is a private/hidden test and not an example test?<br>\nAnd how do we get feedback about the prediction file? I have been submitting Example predictions, but have seen no feedback, such as file-format errors.</p>",
      "rawMarkdown": "Thanks, folks - you advice has been really helpful. Now I want to move on to the real thing ....\n\nThis seems to be the standard instructions:\n\niter_test = env.iter_test() \n(loop start)\n(test_df, sample_prediction_df) = next(iter_test)\n....\nenv.predict(my_prediction_for_test_df)\n(loop end)\n\nHow can we tell that test_df is a private/hidden test and not an example test?\nAnd how do we get feedback about the prediction file? I have been submitting Example predictions, but have seen no feedback, such as file-format errors.",
      "votes": null
    },
    {
      "id": "1053735",
      "postDate": "10/19/2020 09:42:10",
      "content": "<blockquote>\n  <p>How can we tell that test_df is a private/hidden test and not an example test?</p>\n</blockquote>\n<p>kaggle takes care of it when you make a submission…</p>",
      "rawMarkdown": "> How can we tell that test_df is a private/hidden test and not an example test?\n\nkaggle takes care of it when you make a submission...",
      "votes": null
    },
    {
      "id": "1054824",
      "postDate": "10/20/2020 07:38:49",
      "content": "<p>Thanks, Aditya. Looking forward to seeing this in action …. 😊</p>",
      "rawMarkdown": "Thanks, Aditya. Looking forward to seeing this in action .... 😊",
      "votes": null
    },
    {
      "id": "1054870",
      "postDate": "10/20/2020 08:39:55",
      "content": "<p>Here are the next steps:</p>\n<p>1) Have a Notebook that successfully predicts the Example test data when \"Run All\" is clicked<br>\n    Be sure that Setting \"Internet\" is off.<br>\n2) Save to share.<br>\n3) Share with the public<br>\n4) On the Notebooks page, click on it <br>\n5) Execution info<br>\n6) Click on \"Submit\"<br>\nYou will be told that your Notebook is running.</p>",
      "rawMarkdown": "Here are the next steps:\n\n1) Have a Notebook that successfully predicts the Example test data when \"Run All\" is clicked\n    Be sure that Setting \"Internet\" is off.\n2) Save to share.\n3) Share with the public\n4) On the Notebooks page, click on it \n5) Execution info\n6) Click on \"Submit\"\nYou will be told that your Notebook is running.",
      "votes": null
    },
    {
      "id": "1054888",
      "postDate": "10/20/2020 09:08:55",
      "content": "<p>Mike, when you will go to make a submission and run the kernel, kaggle runs it in a private environment and only that environment has real test data (2.5M rows). In the one which we see in the data desc, that's just a sample drawn from it (which happens to be around ~159 records or less). Hope it's clear now.</p>\n<p>Also, you don't have to call <code>next()</code>, you are pretty much given an iterator, just iterate over it to get new rows… (or get chunks of a big dataframe as is the case here)</p>",
      "rawMarkdown": "Mike, when you will go to make a submission and run the kernel, kaggle runs it in a private environment and only that environment has real test data (2.5M rows). In the one which we see in the data desc, that's just a sample drawn from it (which happens to be around ~159 records or less). Hope it's clear now.\n\nAlso, you don't have to call `next()`, you are pretty much given an iterator, just iterate over it to get new rows... (or get chunks of a big dataframe as is the case here)",
      "votes": null
    },
    {
      "id": "1055540",
      "postDate": "10/20/2020 23:13:14",
      "content": "<p>Aditya, thanks. You say: \"when you will go to make a submission and run the kernel, \"</p>\n<p>How do I do this? It seems to be the next step beyond <a href=\"https://www.kaggle.com/sohier/competition-api-detailed-introduction\" target=\"_blank\">https://www.kaggle.com/sohier/competition-api-detailed-introduction</a></p>",
      "rawMarkdown": "Aditya, thanks. You say: \"when you will go to make a submission and run the kernel, \"\n\nHow do I do this? It seems to be the next step beyond https://www.kaggle.com/sohier/competition-api-detailed-introduction",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1053735,
      "author_name": "adityaecdrid",
      "author_url": "",
      "post_date": "10/19/2020 09:42:10",
      "content": "<blockquote>\n  <p>How can we tell that test_df is a private/hidden test and not an example test?</p>\n</blockquote>\n<p>kaggle takes care of it when you make a submission…</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1054824,
      "author_name": "mikel1",
      "author_url": "",
      "post_date": "10/20/2020 07:38:49",
      "content": "<p>Thanks, Aditya. Looking forward to seeing this in action …. 😊</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1054870,
      "author_name": "mikel1",
      "author_url": "",
      "post_date": "10/20/2020 08:39:55",
      "content": "<p>Here are the next steps:</p>\n<p>1) Have a Notebook that successfully predicts the Example test data when \"Run All\" is clicked<br>\n    Be sure that Setting \"Internet\" is off.<br>\n2) Save to share.<br>\n3) Share with the public<br>\n4) On the Notebooks page, click on it <br>\n5) Execution info<br>\n6) Click on \"Submit\"<br>\nYou will be told that your Notebook is running.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1054888,
          "author_name": "adityaecdrid",
          "author_url": "",
          "post_date": "10/20/2020 09:08:55",
          "content": "<p>Mike, when you will go to make a submission and run the kernel, kaggle runs it in a private environment and only that environment has real test data (2.5M rows). In the one which we see in the data desc, that's just a sample drawn from it (which happens to be around ~159 records or less). Hope it's clear now.</p>\n<p>Also, you don't have to call <code>next()</code>, you are pretty much given an iterator, just iterate over it to get new rows… (or get chunks of a big dataframe as is the case here)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1055540,
          "author_name": "mikel1",
          "author_url": "",
          "post_date": "10/20/2020 23:13:14",
          "content": "<p>Aditya, thanks. You say: \"when you will go to make a submission and run the kernel, \"</p>\n<p>How do I do this? It seems to be the next step beyond <a href=\"https://www.kaggle.com/sohier/competition-api-detailed-introduction\" target=\"_blank\">https://www.kaggle.com/sohier/competition-api-detailed-introduction</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1053732": "Thanks, folks - you advice has been really helpful. Now I want to move on to the real thing ....\n\nThis seems to be the standard instructions:\n\niter_test = env.iter_test() \n(loop start)\n(test_df, sample_prediction_df) = next(iter_test)\n....\nenv.predict(my_prediction_for_test_df)\n(loop end)\n\nHow can we tell that test_df is a private/hidden test and not an example test?\nAnd how do we get feedback about the prediction file? I have been submitting Example predictions, but have seen no feedback, such as file-format errors.",
    "1053735": "> How can we tell that test_df is a private/hidden test and not an example test?\n\nkaggle takes care of it when you make a submission...",
    "1054824": "Thanks, Aditya. Looking forward to seeing this in action .... 😊",
    "1054870": "Here are the next steps:\n\n1) Have a Notebook that successfully predicts the Example test data when \"Run All\" is clicked\n    Be sure that Setting \"Internet\" is off.\n2) Save to share.\n3) Share with the public\n4) On the Notebooks page, click on it \n5) Execution info\n6) Click on \"Submit\"\nYou will be told that your Notebook is running.",
    "1054888": "Mike, when you will go to make a submission and run the kernel, kaggle runs it in a private environment and only that environment has real test data (2.5M rows). In the one which we see in the data desc, that's just a sample drawn from it (which happens to be around ~159 records or less). Hope it's clear now.\n\nAlso, you don't have to call `next()`, you are pretty much given an iterator, just iterate over it to get new rows... (or get chunks of a big dataframe as is the case here)",
    "1055540": "Aditya, thanks. You say: \"when you will go to make a submission and run the kernel, \"\n\nHow do I do this? It seems to be the next step beyond https://www.kaggle.com/sohier/competition-api-detailed-introduction"
  },
  "source": "meta"
}