{ "id": 322365, "title": "How well are you prepared for Shakeup...", "url": "/competitions/h-and-m-personalized-fashion-recommendations/discussion/322365", "author_name": "", "post_date": "2022-05-01T18:02:47.411774Z", "votes": 1, "comment_count": 7, "views": 0, "content": "

Although CV and LB looks in line, I have this feeling that 5-15% of test sample in public LB is going to make this competition a perfect candidate for big Shakeup.

", "messages": [ { "id": "1774030", "postDate": "05/01/2022 18:02:47", "content": "

Although CV and LB looks in line, I have this feeling that 5-15% of test sample in public LB is going to make this competition a perfect candidate for big Shakeup.

", "rawMarkdown": "Although CV and LB looks in line, I have this feeling that 5-15% of test sample in public LB is going to make this competition a perfect candidate for big Shakeup.", "votes": null }, { "id": "1774037", "postDate": "05/01/2022 18:14:53", "content": "

After fixing some bugs, the CV and LB is very well corresponded in my case, so I will not expect too big shakeup in this competition. Hope the final result back me up.πŸ˜†

", "rawMarkdown": "After fixing some bugs, the CV and LB is very well corresponded in my case, so I will not expect too big shakeup in this competition. Hope the final result back me up.πŸ˜†", "votes": null }, { "id": "1774061", "postDate": "05/01/2022 18:51:14", "content": "

Personally, as of yet, I haven't been successful at making sense of what's happening in my LGBMRanker training, so I'm working in the dark.

\n

Trying to limit my submissions because of that.
\nBut I'm definitely at risk.

", "rawMarkdown": "Personally, as of yet, I haven't been successful at making sense of what's happening in my LGBMRanker training, so I'm working in the dark.\n\nTrying to limit my submissions because of that.\nBut I'm definitely at risk.", "votes": null }, { "id": "1774065", "postDate": "05/01/2022 19:04:34", "content": "

All the best, hard work gonna pay 🀜
\nStruggling with crossing 0.03 on CV and crunched by managing resource for millions of candidates.

", "rawMarkdown": "All the best, hard work gonna pay 🀜 \nStruggling with crossing 0.03 on CV and crunched by managing resource for millions of candidates.", "votes": null }, { "id": "1774067", "postDate": "05/01/2022 19:06:53", "content": "

In a similar kind of situation
\nIt's a tiring process to debug with all these candidates generation 😣

", "rawMarkdown": "In a similar kind of situation \nIt's a tiring process to debug with all these candidates generation 😣", "votes": null }, { "id": "1774387", "postDate": "05/02/2022 06:20:40", "content": "

This comment is spot on. The competition may well be shaken up, but it's impossible to say for sure at this point. It will be interesting to see how things play out!

", "rawMarkdown": "This comment is spot on. The competition may well be shaken up, but it's impossible to say for sure at this point. It will be interesting to see how things play out!", "votes": null }, { "id": "1780177", "postDate": "05/07/2022 06:54:41", "content": "

In previous tournament experience, 5% of the test sample can really cause a real shakeup in the private LB, for example, the first in the private LB came up from the public LB of ~1500.

", "rawMarkdown": "In previous tournament experience, 5% of the test sample can really cause a real shakeup in the private LB, for example, the first in the private LB came up from the public LB of ~1500.", "votes": null }, { "id": "1780231", "postDate": "05/07/2022 08:08:16", "content": "

People have reported a difference of 0.004-0.006 difference in their CV and LB on this 5-15% of data, judging by that, we can expect a difference of 0.035-0.050 between CV and LB on the private testset.

\n

Shakeup is going to be imminent… πŸ‘€

", "rawMarkdown": "People have reported a difference of 0.004-0.006 difference in their CV and LB on this 5-15% of data, judging by that, we can expect a difference of 0.035-0.050 between CV and LB on the private testset. \n\nShakeup is going to be imminent... πŸ‘€", "votes": null } ], "comments": [ { "id": 1774037, "author_name": "lihaorocky", "author_url": "", "post_date": "05/01/2022 18:14:53", "content": "

After fixing some bugs, the CV and LB is very well corresponded in my case, so I will not expect too big shakeup in this competition. Hope the final result back me up.πŸ˜†

", "votes": null, "replies": [ { "id": 1774065, "author_name": "tarique7", "author_url": "", "post_date": "05/01/2022 19:04:34", "content": "

All the best, hard work gonna pay 🀜
\nStruggling with crossing 0.03 on CV and crunched by managing resource for millions of candidates.

", "votes": null, "replies": [] } ] }, { "id": 1774061, "author_name": "jacob34", "author_url": "", "post_date": "05/01/2022 18:51:14", "content": "

Personally, as of yet, I haven't been successful at making sense of what's happening in my LGBMRanker training, so I'm working in the dark.

\n

Trying to limit my submissions because of that.
\nBut I'm definitely at risk.

", "votes": null, "replies": [ { "id": 1774067, "author_name": "tarique7", "author_url": "", "post_date": "05/01/2022 19:06:53", "content": "

In a similar kind of situation
\nIt's a tiring process to debug with all these candidates generation 😣

", "votes": null, "replies": [] } ] }, { "id": 1774387, "author_name": "", "author_url": "", "post_date": "05/02/2022 06:20:40", "content": "

This comment is spot on. The competition may well be shaken up, but it's impossible to say for sure at this point. It will be interesting to see how things play out!

", "votes": null, "replies": [] }, { "id": 1780177, "author_name": "lemelo", "author_url": "", "post_date": "05/07/2022 06:54:41", "content": "

In previous tournament experience, 5% of the test sample can really cause a real shakeup in the private LB, for example, the first in the private LB came up from the public LB of ~1500.

", "votes": null, "replies": [ { "id": 1780231, "author_name": "tarique7", "author_url": "", "post_date": "05/07/2022 08:08:16", "content": "

People have reported a difference of 0.004-0.006 difference in their CV and LB on this 5-15% of data, judging by that, we can expect a difference of 0.035-0.050 between CV and LB on the private testset.

\n

Shakeup is going to be imminent… πŸ‘€

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1774030": "Although CV and LB looks in line, I have this feeling that 5-15% of test sample in public LB is going to make this competition a perfect candidate for big Shakeup.", "1774037": "After fixing some bugs, the CV and LB is very well corresponded in my case, so I will not expect too big shakeup in this competition. Hope the final result back me up.πŸ˜†", "1774061": "Personally, as of yet, I haven't been successful at making sense of what's happening in my LGBMRanker training, so I'm working in the dark.\n\nTrying to limit my submissions because of that.\nBut I'm definitely at risk.", "1774065": "All the best, hard work gonna pay 🀜 \nStruggling with crossing 0.03 on CV and crunched by managing resource for millions of candidates.", "1774067": "In a similar kind of situation \nIt's a tiring process to debug with all these candidates generation 😣", "1774387": "This comment is spot on. The competition may well be shaken up, but it's impossible to say for sure at this point. It will be interesting to see how things play out!", "1780177": "In previous tournament experience, 5% of the test sample can really cause a real shakeup in the private LB, for example, the first in the private LB came up from the public LB of ~1500.", "1780231": "People have reported a difference of 0.004-0.006 difference in their CV and LB on this 5-15% of data, judging by that, we can expect a difference of 0.035-0.050 between CV and LB on the private testset. \n\nShakeup is going to be imminent... πŸ‘€" }, "source": "meta" }