Just to be sure, in the end when our notebooks are scored for the private leaderboard, it will provide our notebook for exactly the same name and format of data, and expect to output exactly the same name and format of data?
\nThere is one of my submissions that I miswrote the column names and was too late, but the submission still scored fine. Is it also how private submission will behave?
\nAlso, for the private dataset, does it have a similar skewed distribution as the public?
", "messages": [ { "id": "3027319", "postDate": "10/24/2024 17:12:46", "content": "Just to be sure, in the end when our notebooks are scored for the private leaderboard, it will provide our notebook for exactly the same name and format of data, and expect to output exactly the same name and format of data?
\nThere is one of my submissions that I miswrote the column names and was too late, but the submission still scored fine. Is it also how private submission will behave?
\nAlso, for the private dataset, does it have a similar skewed distribution as the public?
", "rawMarkdown": "Just to be sure, in the end when our notebooks are scored for the private leaderboard, it will provide our notebook for exactly the same name and format of data, and expect to output exactly the same name and format of data? \n\nThere is one of my submissions that I miswrote the column names and was too late, but the submission still scored fine. Is it also how private submission will behave?\n\nAlso, for the private dataset, does it have a similar skewed distribution as the public?", "votes": null }, { "id": "3030811", "postDate": "10/28/2024 23:12:15", "content": "You do not need to provide any notebooks or code along with your submission to the leaderboard, you just need to submit your predictions. After you have made multiple submissions there should be a tab on the Kaggle site called \"submissions\", from there you can see the public scores assocaited with each of your submissions. The custom label you used when submitting is only for you to track and organize your submissions, such as V1, V2, test1, test2. You can put a checkmark next to two submissions that you want to count for your private score once the competition ends. If you have trouble we can schedule a zoom call on Thursday/Friday to clear any confusion.
\nI can confirm that the test dataset has as similar skewed distribution similar to the training data. I used a binned stratified train/test split to ensure consistency between both distributions.
", "rawMarkdown": "You do not need to provide any notebooks or code along with your submission to the leaderboard, you just need to submit your predictions. After you have made multiple submissions there should be a tab on the Kaggle site called \"submissions\", from there you can see the public scores assocaited with each of your submissions. The custom label you used when submitting is only for you to track and organize your submissions, such as V1, V2, test1, test2. You can put a checkmark next to two submissions that you want to count for your private score once the competition ends. If you have trouble we can schedule a zoom call on Thursday/Friday to clear any confusion. \n\nI can confirm that the test dataset has as similar skewed distribution similar to the training data. I used a binned stratified train/test split to ensure consistency between both distributions.", "votes": null } ], "comments": [ { "id": 3030811, "author_name": "robertjirsaraie", "author_url": "", "post_date": "10/28/2024 23:12:15", "content": "You do not need to provide any notebooks or code along with your submission to the leaderboard, you just need to submit your predictions. After you have made multiple submissions there should be a tab on the Kaggle site called \"submissions\", from there you can see the public scores assocaited with each of your submissions. The custom label you used when submitting is only for you to track and organize your submissions, such as V1, V2, test1, test2. You can put a checkmark next to two submissions that you want to count for your private score once the competition ends. If you have trouble we can schedule a zoom call on Thursday/Friday to clear any confusion.
\nI can confirm that the test dataset has as similar skewed distribution similar to the training data. I used a binned stratified train/test split to ensure consistency between both distributions.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "3027319": "Just to be sure, in the end when our notebooks are scored for the private leaderboard, it will provide our notebook for exactly the same name and format of data, and expect to output exactly the same name and format of data? \n\nThere is one of my submissions that I miswrote the column names and was too late, but the submission still scored fine. Is it also how private submission will behave?\n\nAlso, for the private dataset, does it have a similar skewed distribution as the public?", "3030811": "You do not need to provide any notebooks or code along with your submission to the leaderboard, you just need to submit your predictions. After you have made multiple submissions there should be a tab on the Kaggle site called \"submissions\", from there you can see the public scores assocaited with each of your submissions. The custom label you used when submitting is only for you to track and organize your submissions, such as V1, V2, test1, test2. You can put a checkmark next to two submissions that you want to count for your private score once the competition ends. If you have trouble we can schedule a zoom call on Thursday/Friday to clear any confusion. \n\nI can confirm that the test dataset has as similar skewed distribution similar to the training data. I used a binned stratified train/test split to ensure consistency between both distributions." }, "source": "meta" }