{
  "id": 180360,
  "title": "Not getting test_audio even after submission???",
  "url": "/competitions/birdsong-recognition/discussion/180360",
  "author_name": "",
  "post_date": "2020-09-04T18:06:38.106531600Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>No such file or directory: '/kaggle/input/birdsong-recognition/test_audio/0a997dff022e3ad9744d4e7bbf923288.mp3'</p>\n<p>I continually get this error. I realize test data is not available in the environment, but even when i submit I am not locating the test audio data. wth???</p>",
  "messages": [
    {
      "id": "998425",
      "postDate": "09/04/2020 18:06:38",
      "content": "<p>No such file or directory: '/kaggle/input/birdsong-recognition/test_audio/0a997dff022e3ad9744d4e7bbf923288.mp3'</p>\n<p>I continually get this error. I realize test data is not available in the environment, but even when i submit I am not locating the test audio data. wth???</p>",
      "rawMarkdown": "No such file or directory: '/kaggle/input/birdsong-recognition/test_audio/0a997dff022e3ad9744d4e7bbf923288.mp3'\n\nI continually get this error. I realize test data is not available in the environment, but even when i submit I am not locating the test audio data. wth???",
      "votes": null
    },
    {
      "id": "998579",
      "postDate": "09/04/2020 20:54:22",
      "content": "<p>You have to use e.g. birdcall-check to be able to submit on the \"real\" test-data there are notebooks for that e.g. <a href=\"https://www.kaggle.com/shonenkov/sample-submission-using-custom-check\" target=\"_blank\">https://www.kaggle.com/shonenkov/sample-submission-using-custom-check</a></p>\n<p>The idea is creating a submission on birdcall-check and when it reruns the code during submission it takes a different path and uses the \"real\" test-data.</p>",
      "rawMarkdown": "You have to use e.g. birdcall-check to be able to submit on the \"real\" test-data there are notebooks for that e.g. https://www.kaggle.com/shonenkov/sample-submission-using-custom-check\n\nThe idea is creating a submission on birdcall-check and when it reruns the code during submission it takes a different path and uses the \"real\" test-data.",
      "votes": null
    },
    {
      "id": "998701",
      "postDate": "09/05/2020 01:28:49",
      "content": "<p>Save &gt; Save and Run All &gt; Commit : does NOT see real test data</p>\n<p>Save &gt; Save and Run All &gt; Commit &gt; (wait for notebook to run) &gt; Open saved notebook &gt; Submit to Competition &gt; DOES see real test data in rerunning your code.</p>\n<p>You don't ever get to see notebook results on actual test data in order to keep it 100% locked away. The code from your submitted notebook is rerun using the actual data. Makes it extremely hard to set up and debug (but I do understand why the organisers have done this - it defeats the purpose of the comp if people can probe the results from running models on the data in depth).</p>",
      "rawMarkdown": "Save > Save and Run All > Commit : does NOT see real test data\n\nSave > Save and Run All > Commit > (wait for notebook to run) > Open saved notebook > Submit to Competition > DOES see real test data in rerunning your code.\n\nYou don't ever get to see notebook results on actual test data in order to keep it 100% locked away. The code from your submitted notebook is rerun using the actual data. Makes it extremely hard to set up and debug (but I do understand why the organisers have done this - it defeats the purpose of the comp if people can probe the results from running models on the data in depth).",
      "votes": null
    },
    {
      "id": "999420",
      "postDate": "09/05/2020 16:20:56",
      "content": "<p>This works in R:<br>\nif (!dir.exists('/kaggle/input/birdsong-recognition/test_audio')){<br>\n sub1&lt;- read_csv('/kaggle/input/birdsong-recognition/sample_submission.csv')<br>\n write.csv(sub1, \"submission.csv\", row.names = F)<br>\n quit(save = \"no\")<br>\n}</p>\n<p>Thanks for the procedure.</p>",
      "rawMarkdown": "This works in R:\n\nif (!dir.exists('/kaggle/input/birdsong-recognition/test_audio')){\n  sub1<- read_csv('/kaggle/input/birdsong-recognition/sample_submission.csv')\n  write.csv(sub1, \"submission.csv\", row.names = F)\n  quit(save = \"no\")\n}\n \nThanks for the procedure.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 998579,
      "author_name": "aliabdin1",
      "author_url": "",
      "post_date": "09/04/2020 20:54:22",
      "content": "<p>You have to use e.g. birdcall-check to be able to submit on the \"real\" test-data there are notebooks for that e.g. <a href=\"https://www.kaggle.com/shonenkov/sample-submission-using-custom-check\" target=\"_blank\">https://www.kaggle.com/shonenkov/sample-submission-using-custom-check</a></p>\n<p>The idea is creating a submission on birdcall-check and when it reruns the code during submission it takes a different path and uses the \"real\" test-data.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 998701,
      "author_name": "davidedwards1",
      "author_url": "",
      "post_date": "09/05/2020 01:28:49",
      "content": "<p>Save &gt; Save and Run All &gt; Commit : does NOT see real test data</p>\n<p>Save &gt; Save and Run All &gt; Commit &gt; (wait for notebook to run) &gt; Open saved notebook &gt; Submit to Competition &gt; DOES see real test data in rerunning your code.</p>\n<p>You don't ever get to see notebook results on actual test data in order to keep it 100% locked away. The code from your submitted notebook is rerun using the actual data. Makes it extremely hard to set up and debug (but I do understand why the organisers have done this - it defeats the purpose of the comp if people can probe the results from running models on the data in depth).</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 999420,
      "author_name": "tomkkk",
      "author_url": "",
      "post_date": "09/05/2020 16:20:56",
      "content": "<p>This works in R:<br>\nif (!dir.exists('/kaggle/input/birdsong-recognition/test_audio')){<br>\n sub1&lt;- read_csv('/kaggle/input/birdsong-recognition/sample_submission.csv')<br>\n write.csv(sub1, \"submission.csv\", row.names = F)<br>\n quit(save = \"no\")<br>\n}</p>\n<p>Thanks for the procedure.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "998425": "No such file or directory: '/kaggle/input/birdsong-recognition/test_audio/0a997dff022e3ad9744d4e7bbf923288.mp3'\n\nI continually get this error. I realize test data is not available in the environment, but even when i submit I am not locating the test audio data. wth???",
    "998579": "You have to use e.g. birdcall-check to be able to submit on the \"real\" test-data there are notebooks for that e.g. https://www.kaggle.com/shonenkov/sample-submission-using-custom-check\n\nThe idea is creating a submission on birdcall-check and when it reruns the code during submission it takes a different path and uses the \"real\" test-data.",
    "998701": "Save > Save and Run All > Commit : does NOT see real test data\n\nSave > Save and Run All > Commit > (wait for notebook to run) > Open saved notebook > Submit to Competition > DOES see real test data in rerunning your code.\n\nYou don't ever get to see notebook results on actual test data in order to keep it 100% locked away. The code from your submitted notebook is rerun using the actual data. Makes it extremely hard to set up and debug (but I do understand why the organisers have done this - it defeats the purpose of the comp if people can probe the results from running models on the data in depth).",
    "999420": "This works in R:\n\nif (!dir.exists('/kaggle/input/birdsong-recognition/test_audio')){\n  sub1<- read_csv('/kaggle/input/birdsong-recognition/sample_submission.csv')\n  write.csv(sub1, \"submission.csv\", row.names = F)\n  quit(save = \"no\")\n}\n \nThanks for the procedure."
  },
  "source": "meta"
}