The private leaderboard tab says: The private leaderboard is calculated over the same rows as the public leaderboard in this competition. Does this mean that we are getting the same scores in private leaderboard so my current rank would be the final one?
", "messages": [ { "id": "792287", "postDate": "03/31/2020 04:25:20", "content": "The private leaderboard tab says: The private leaderboard is calculated over the same rows as the public leaderboard in this competition. Does this mean that we are getting the same scores in private leaderboard so my current rank would be the final one?
", "rawMarkdown": "The private leaderboard tab says: The private leaderboard is calculated over the same rows as the public leaderboard in this competition. Does this mean that we are getting the same scores in private leaderboard so my current rank would be the final one?", "votes": null }, { "id": "792307", "postDate": "03/31/2020 04:45:11", "content": "March 3, 2020 - Entry deadline. You must accept the competition rules before this date in order to compete.
\n\nMarch 3, 2020 - Team Merger deadline. This is the last day participants may join or merge teams.
\n\nMarch 31, 2020 - Final submission deadline.
\n\n```\n(Estimated) April 22, 2020 - Private leaderboard revealed.
\n\nNote: Because the private test set is held outside of Kaggle's platform, the private leaderboard will not be revealed immediately. We have estimated the scoring timeframe, but this is subject to revision.\n```
\n\nKaggle have public and private dataset same. But private LB scores calculated with data owned by DFDC team
", "rawMarkdown": "March 3, 2020 - Entry deadline. You must accept the competition rules before this date in order to compete.\n\nMarch 3, 2020 - Team Merger deadline. This is the last day participants may join or merge teams.\n\nMarch 31, 2020 - Final submission deadline.\n\n```\n(Estimated) April 22, 2020 - Private leaderboard revealed.\n\nNote: Because the private test set is held outside of Kaggle's platform, the private leaderboard will not be revealed immediately. We have estimated the scoring timeframe, but this is subject to revision.\n```\n\nKaggle have public and private dataset same. But private LB scores calculated with data owned by DFDC team", "votes": null }, { "id": "792963", "postDate": "03/31/2020 17:25:32", "content": "Sorry, but your words are confusing. Kaggle does not have the private dataset.\nAnd our final scores will be calculated on the data owned by DFDC team, not Kaggle.
", "rawMarkdown": "Sorry, but your words are confusing. **Kaggle does not have the private dataset.**\nAnd our final scores will be calculated on the data owned by DFDC team, not Kaggle.", "votes": null }, { "id": "793045", "postDate": "03/31/2020 18:54:56", "content": "@seshurajup is correct. The private leaderboard will be evaluated by the host re-running participants' selected code against a fully withheld private test set. Because this takes time, the reveal of the private leaderboard scoring is not expected to be revealed until ~3 weeks following the submission deadline. So although the current private leaderboard may indicate it will be calculated over the same rows, this is not the case for the true and final private leaderboard.
", "rawMarkdown": "seshurajup is correct. The private leaderboard will be evaluated by the host re-running participants' selected code against a fully withheld private test set. Because this takes time, the reveal of the private leaderboard scoring is not expected to be revealed until ~3 weeks following the submission deadline. So although the current private leaderboard may indicate it will be calculated over the same rows, this is not the case for the true and final private leaderboard.", "votes": null }, { "id": "793165", "postDate": "03/31/2020 20:59:46", "content": "The private dataset is exactly 10x of the public one.
", "rawMarkdown": "The private dataset is exactly 10x of the public one.", "votes": null }, { "id": "793288", "postDate": "03/31/2020 23:31:10", "content": "How will be the code used by the host? Will they use the Docker images + versioned datasets, e.g. a snapshot of what we had while we had a successful submision?? Also if we have a successful submission and selection does it guarantee that there won't be any potential problems during inference outside the Kaggle platform? For example, if a public dataset we depend on the Kaggle platform happens to be deleted or modified would it affect anything?
", "rawMarkdown": "How will be the code used by the host? Will they use the Docker images + versioned datasets, e.g. a snapshot of what we had while we had a successful submision?? Also if we have a successful submission and selection does it guarantee that there won't be any potential problems during inference outside the Kaggle platform? For example, if a public dataset we depend on the Kaggle platform happens to be deleted or modified would it affect anything?", "votes": null } ], "comments": [ { "id": 792307, "author_name": "seshurajup", "author_url": "", "post_date": "03/31/2020 04:45:11", "content": "March 3, 2020 - Entry deadline. You must accept the competition rules before this date in order to compete.
\n\nMarch 3, 2020 - Team Merger deadline. This is the last day participants may join or merge teams.
\n\nMarch 31, 2020 - Final submission deadline.
\n\n```\n(Estimated) April 22, 2020 - Private leaderboard revealed.
\n\nNote: Because the private test set is held outside of Kaggle's platform, the private leaderboard will not be revealed immediately. We have estimated the scoring timeframe, but this is subject to revision.\n```
\n\nKaggle have public and private dataset same. But private LB scores calculated with data owned by DFDC team
", "votes": null, "replies": [ { "id": 792963, "author_name": "vpaslay", "author_url": "", "post_date": "03/31/2020 17:25:32", "content": "Sorry, but your words are confusing. Kaggle does not have the private dataset.\nAnd our final scores will be calculated on the data owned by DFDC team, not Kaggle.
", "votes": null, "replies": [] }, { "id": 793045, "author_name": "juliaelliott", "author_url": "", "post_date": "03/31/2020 18:54:56", "content": "@seshurajup is correct. The private leaderboard will be evaluated by the host re-running participants' selected code against a fully withheld private test set. Because this takes time, the reveal of the private leaderboard scoring is not expected to be revealed until ~3 weeks following the submission deadline. So although the current private leaderboard may indicate it will be calculated over the same rows, this is not the case for the true and final private leaderboard.
", "votes": null, "replies": [] }, { "id": 793288, "author_name": "keremt", "author_url": "", "post_date": "03/31/2020 23:31:10", "content": "How will be the code used by the host? Will they use the Docker images + versioned datasets, e.g. a snapshot of what we had while we had a successful submision?? Also if we have a successful submission and selection does it guarantee that there won't be any potential problems during inference outside the Kaggle platform? For example, if a public dataset we depend on the Kaggle platform happens to be deleted or modified would it affect anything?
", "votes": null, "replies": [] } ] }, { "id": 793165, "author_name": "", "author_url": "", "post_date": "03/31/2020 20:59:46", "content": "The private dataset is exactly 10x of the public one.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "792287": "The private leaderboard tab says: The private leaderboard is calculated over the same rows as the public leaderboard in this competition. Does this mean that we are getting the same scores in private leaderboard so my current rank would be the final one?", "792307": "March 3, 2020 - Entry deadline. You must accept the competition rules before this date in order to compete.\n\nMarch 3, 2020 - Team Merger deadline. This is the last day participants may join or merge teams.\n\nMarch 31, 2020 - Final submission deadline.\n\n```\n(Estimated) April 22, 2020 - Private leaderboard revealed.\n\nNote: Because the private test set is held outside of Kaggle's platform, the private leaderboard will not be revealed immediately. We have estimated the scoring timeframe, but this is subject to revision.\n```\n\nKaggle have public and private dataset same. But private LB scores calculated with data owned by DFDC team", "792963": "Sorry, but your words are confusing. **Kaggle does not have the private dataset.**\nAnd our final scores will be calculated on the data owned by DFDC team, not Kaggle.", "793045": "seshurajup is correct. The private leaderboard will be evaluated by the host re-running participants' selected code against a fully withheld private test set. Because this takes time, the reveal of the private leaderboard scoring is not expected to be revealed until ~3 weeks following the submission deadline. So although the current private leaderboard may indicate it will be calculated over the same rows, this is not the case for the true and final private leaderboard.", "793165": "The private dataset is exactly 10x of the public one.", "793288": "How will be the code used by the host? Will they use the Docker images + versioned datasets, e.g. a snapshot of what we had while we had a successful submision?? Also if we have a successful submission and selection does it guarantee that there won't be any potential problems during inference outside the Kaggle platform? For example, if a public dataset we depend on the Kaggle platform happens to be deleted or modified would it affect anything?" }, "source": "meta" }