{
  "id": 206355,
  "title": "Does the notebook I submit also run on the private subset of the test data?",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/206355",
  "author_name": "",
  "post_date": "2020-12-24T07:49:19.584645800Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Since this is my first competition I am still a bit confused by all of these things. <br>\nMy notebook takes somewhere over 6 hours to run when I submit it now. So does that mean that it is running only on the data used for this public leaderboard or it's running on both the public and private data?</p>",
  "messages": [
    {
      "id": "1124783",
      "postDate": "12/24/2020 07:49:19",
      "content": "<p>Since this is my first competition I am still a bit confused by all of these things. <br>\nMy notebook takes somewhere over 6 hours to run when I submit it now. So does that mean that it is running only on the data used for this public leaderboard or it's running on both the public and private data?</p>",
      "rawMarkdown": "Since this is my first competition I am still a bit confused by all of these things. \nMy notebook takes somewhere over 6 hours to run when I submit it now. So does that mean that it is running only on the data used for this public leaderboard or it's running on both the public and private data?",
      "votes": null
    },
    {
      "id": "1124953",
      "postDate": "12/24/2020 09:40:19",
      "content": "<p>When you make the first Save/Submit the notebook runs only the single image listed in the test file that your can read or download - stage 1.  The \"public\" data is this single file.  </p>\n<p>After that is competed - click to preview your created submission.csv - it should only show one image.  Immediately - Click on Submit and Kaggle will not run the notebook with the full test set - they will report back (some hours later) the results - stage 2.  Go ahead and click on the View Submissions.</p>\n<p>Your model has than processed the entire test set - but they only report back on 31% of the images processed.  I think 15,000 is the number of images for the private data.</p>\n<p>Your question a bit confusing - 6 hours in which of the stages mentioned above.</p>",
      "rawMarkdown": "When you make the first Save/Submit the notebook runs only the single image listed in the test file that your can read or download - stage 1.  The \"public\" data is this single file.  \n\nAfter that is competed - click to preview your created submission.csv - it should only show one image.  Immediately - Click on Submit and Kaggle will not run the notebook with the full test set - they will report back (some hours later) the results - stage 2.  Go ahead and click on the View Submissions.\n\nYour model has than processed the entire test set - but they only report back on 31% of the images processed.  I think 15,000 is the number of images for the private data.\n\nYour question a bit confusing - 6 hours in which of the stages mentioned above.",
      "votes": null
    },
    {
      "id": "1125151",
      "postDate": "12/24/2020 12:23:03",
      "content": "<p>6 hours after submitting to the competition. I wanted to know if it runs on 31% of the 15000 images or on the entire dataset. Thanks for informing that it runs on the entire data. Otherwise I would have to reduce the run time for the inference notebook.</p>",
      "rawMarkdown": "6 hours after submitting to the competition. I wanted to know if it runs on 31% of the 15000 images or on the entire dataset. Thanks for informing that it runs on the entire data. Otherwise I would have to reduce the run time for the inference notebook.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1124953,
      "author_name": "pcjimmmy",
      "author_url": "",
      "post_date": "12/24/2020 09:40:19",
      "content": "<p>When you make the first Save/Submit the notebook runs only the single image listed in the test file that your can read or download - stage 1.  The \"public\" data is this single file.  </p>\n<p>After that is competed - click to preview your created submission.csv - it should only show one image.  Immediately - Click on Submit and Kaggle will not run the notebook with the full test set - they will report back (some hours later) the results - stage 2.  Go ahead and click on the View Submissions.</p>\n<p>Your model has than processed the entire test set - but they only report back on 31% of the images processed.  I think 15,000 is the number of images for the private data.</p>\n<p>Your question a bit confusing - 6 hours in which of the stages mentioned above.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1125151,
          "author_name": "ashishsingh226",
          "author_url": "",
          "post_date": "12/24/2020 12:23:03",
          "content": "<p>6 hours after submitting to the competition. I wanted to know if it runs on 31% of the 15000 images or on the entire dataset. Thanks for informing that it runs on the entire data. Otherwise I would have to reduce the run time for the inference notebook.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1124783": "Since this is my first competition I am still a bit confused by all of these things. \nMy notebook takes somewhere over 6 hours to run when I submit it now. So does that mean that it is running only on the data used for this public leaderboard or it's running on both the public and private data?",
    "1124953": "When you make the first Save/Submit the notebook runs only the single image listed in the test file that your can read or download - stage 1.  The \"public\" data is this single file.  \n\nAfter that is competed - click to preview your created submission.csv - it should only show one image.  Immediately - Click on Submit and Kaggle will not run the notebook with the full test set - they will report back (some hours later) the results - stage 2.  Go ahead and click on the View Submissions.\n\nYour model has than processed the entire test set - but they only report back on 31% of the images processed.  I think 15,000 is the number of images for the private data.\n\nYour question a bit confusing - 6 hours in which of the stages mentioned above.",
    "1125151": "6 hours after submitting to the competition. I wanted to know if it runs on 31% of the 15000 images or on the entire dataset. Thanks for informing that it runs on the entire data. Otherwise I would have to reduce the run time for the inference notebook."
  },
  "source": "meta"
}