Dear Hosts, dear Kaggle Team, dear Wheat-Finders ..
\n\nhow is the cleaning of the Leaderboard of non-compliant submissions/solutions done?\nWhen will it be concluded?
\n\nI do not expect any Kaggler to use non-compliant submissions on purpose, but it still will have happend a lot of times and the great Shakeup (+-350 places happens quite frequently) shows it.
\n\nGuess there will be an own competition:\nGiven: a notebook name, a FORKED FROM name, and all the code:\nCan you predict if the submission is compliant or YOLOV5?;)?!
\n\n@tanulsingh077, @frtgnn @harshris21, @rohandx1996
", "messages": [ { "id": 958903, "postDate": "2020-08-05T07:55:49.590Z", "content": "Dear Hosts, dear Kaggle Team, dear Wheat-Finders ..
\n\nhow is the cleaning of the Leaderboard of non-compliant submissions/solutions done?\nWhen will it be concluded?
\n\nI do not expect any Kaggler to use non-compliant submissions on purpose, but it still will have happend a lot of times and the great Shakeup (+-350 places happens quite frequently) shows it.
\n\nGuess there will be an own competition:\nGiven: a notebook name, a FORKED FROM name, and all the code:\nCan you predict if the submission is compliant or YOLOV5?;)?!
\n\n@tanulsingh077, @frtgnn @harshris21, @rohandx1996
", "rawMarkdown": "Dear Hosts, dear Kaggle Team, dear Wheat-Finders ..\n\nhow is the cleaning of the Leaderboard of non-compliant submissions/solutions done?\nWhen will it be concluded?\n\nI do not expect any Kaggler to use non-compliant submissions on purpose, but it still will have happend a lot of times and the great Shakeup (+-350 places happens quite frequently) shows it.\n\n- A lot of teams are **not following the discussions** and many people are not aware that Yolov5 is not compliant. \n- Some teams may have chosen a yolov5 based submission **by mistake/accidentally**\n- The **Auto-Pick of solutions** (if you don't pick manually) will take the best public KB score, no matter if they are compliant.\n\nGuess there will be an own competition:\nGiven: a notebook name, a FORKED FROM name, and all the code:\nCan you predict if the submission is compliant or YOLOV5?;)?!\n\n@tanulsingh077, @frtgnn @harshris21, @rohandx1996 ", "votes": 6 }, { "id": 960440, "postDate": "2020-08-06T12:05:44.200Z", "content": "Hm, sadly no update from the hosts/kaggle in any thread, oder could anyone of you find something?
", "rawMarkdown": "Hm, sadly no update from the hosts/kaggle in any thread, oder could anyone of you find something?\n", "replies": [ { "id": 960463, "postDate": "2020-08-06T12:24:54.120Z", "content": "What is your wonder and question, maybe I can help. If it's about the non-compliant submissions and reviewing, I think we just must wait and see, much work for them right now handling the team within the medalzon. But first they will re-run the private test-set, so I think it will be a two-step action. As to how, what,why questions, it's better to see the result, ta many combinations of the outcome, let them handle it first, then we see, and the questions will become more obvious to ask, also some will automatically be answered due to the outcome.
", "rawMarkdown": "What is your wonder and question, maybe I can help. If it's about the non-compliant submissions and reviewing, I think we just must wait and see, much work for them right now handling the team within the medalzon. But first they will re-run the private test-set, so I think it will be a two-step action. As to how, what,why questions, it's better to see the result, ta many combinations of the outcome, let them handle it first, then we see, and the questions will become more obvious to ask, also some will automatically be answered due to the outcome." }, { "id": 960474, "postDate": "2020-08-06T12:34:49.017Z", "content": "I also saw that you have a great start after just 3 month 👍 and welcome to Kaggle. \nI can tell you that it is usually not this complicated, and I think we make it more complicated, we know which models can be used, then if the handling is like the previous competition with similar rules, we have probably seen a standard on how it is handled without to get it in writing. So we'll probably just wait and see.
", "rawMarkdown": "I also saw that you have a great start after just 3 month 👍 and welcome to Kaggle. \nI can tell you that it is usually not this complicated, and I think we make it more complicated, we know which models can be used, then if the handling is like the previous competition with similar rules, we have probably seen a standard on how it is handled without to get it in writing. So we'll probably just wait and see.", "votes": 1 } ] }, { "id": 958963, "postDate": "2020-08-05T08:39:21.483Z", "content": "With all due respect to those who are new and are among their first competitions, anything can happen and not easy to sort through all the terminologies and information,I know, we all know, we have all been there.
\n\nHaving said that, that can be some of the reasons you mentioned but I think those worked actively in this competions can't have missed all stop signs, written all over the place, in the forums, kernels and even as the top kernels names.\nI think the majority can be those not worked actively and maybe forked top Yolo kernels or used thier own Yolo model in the beginning or some point after and left it there, those kernels beats most of the other not-Yolo kernels that have severals of weeks of work all the way to the deadline, and then the final auto-pick algorithm handled the rest.
", "rawMarkdown": "With all due respect to those who are new and are among their first competitions, anything can happen and not easy to sort through all the terminologies and information,I know, we all know, we have all been there.\n\nHaving said that, that can be some of the reasons you mentioned but I think those worked actively in this competions can't have missed all stop signs, written all over the place, in the forums, kernels and even as the top kernels names.\nI think the majority can be those not worked actively and maybe forked top Yolo kernels or used thier own Yolo model in the beginning or some point after and left it there, those kernels beats most of the other not-Yolo kernels that have severals of weeks of work all the way to the deadline, and then the final auto-pick algorithm handled the rest.", "replies": [ { "id": 958966, "postDate": "2020-08-05T08:44:57.473Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 958938, "postDate": "2020-08-05T08:23:23.983Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 960440, "author_name": "from coffee import *", "author_url": "", "post_date": "2020-08-06T12:05:44.200000", "content": "Hm, sadly no update from the hosts/kaggle in any thread, oder could anyone of you find something?
", "votes": 0, "replies": [ { "id": 960463, "author_name": "Kirderf", "author_url": "", "post_date": "2020-08-06T12:24:54.120000", "content": "What is your wonder and question, maybe I can help. If it's about the non-compliant submissions and reviewing, I think we just must wait and see, much work for them right now handling the team within the medalzon. But first they will re-run the private test-set, so I think it will be a two-step action. As to how, what,why questions, it's better to see the result, ta many combinations of the outcome, let them handle it first, then we see, and the questions will become more obvious to ask, also some will automatically be answered due to the outcome.
", "votes": 0, "replies": [] }, { "id": 960474, "author_name": "Kirderf", "author_url": "", "post_date": "2020-08-06T12:34:49.017000", "content": "I also saw that you have a great start after just 3 month 👍 and welcome to Kaggle. \nI can tell you that it is usually not this complicated, and I think we make it more complicated, we know which models can be used, then if the handling is like the previous competition with similar rules, we have probably seen a standard on how it is handled without to get it in writing. So we'll probably just wait and see.
", "votes": 1, "replies": [] } ] }, { "id": 958963, "author_name": "Kirderf", "author_url": "", "post_date": "2020-08-05T08:39:21.483000", "content": "With all due respect to those who are new and are among their first competitions, anything can happen and not easy to sort through all the terminologies and information,I know, we all know, we have all been there.
\n\nHaving said that, that can be some of the reasons you mentioned but I think those worked actively in this competions can't have missed all stop signs, written all over the place, in the forums, kernels and even as the top kernels names.\nI think the majority can be those not worked actively and maybe forked top Yolo kernels or used thier own Yolo model in the beginning or some point after and left it there, those kernels beats most of the other not-Yolo kernels that have severals of weeks of work all the way to the deadline, and then the final auto-pick algorithm handled the rest.
", "votes": 0, "replies": [ { "id": 958966, "author_name": "", "author_url": "", "post_date": "2020-08-05T08:44:57.473000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 958938, "author_name": "", "author_url": "", "post_date": "2020-08-05T08:23:23.983000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "958903": "Dear Hosts, dear Kaggle Team, dear Wheat-Finders ..\n\nhow is the cleaning of the Leaderboard of non-compliant submissions/solutions done?\nWhen will it be concluded?\n\nI do not expect any Kaggler to use non-compliant submissions on purpose, but it still will have happend a lot of times and the great Shakeup (+-350 places happens quite frequently) shows it.\n\n- A lot of teams are **not following the discussions** and many people are not aware that Yolov5 is not compliant. \n- Some teams may have chosen a yolov5 based submission **by mistake/accidentally**\n- The **Auto-Pick of solutions** (if you don't pick manually) will take the best public KB score, no matter if they are compliant.\n\nGuess there will be an own competition:\nGiven: a notebook name, a FORKED FROM name, and all the code:\nCan you predict if the submission is compliant or YOLOV5?;)?!\n\n@tanulsingh077, @frtgnn @harshris21, @rohandx1996 ", "960440": "Hm, sadly no update from the hosts/kaggle in any thread, oder could anyone of you find something?\n", "958963": "With all due respect to those who are new and are among their first competitions, anything can happen and not easy to sort through all the terminologies and information,I know, we all know, we have all been there.\n\nHaving said that, that can be some of the reasons you mentioned but I think those worked actively in this competions can't have missed all stop signs, written all over the place, in the forums, kernels and even as the top kernels names.\nI think the majority can be those not worked actively and maybe forked top Yolo kernels or used thier own Yolo model in the beginning or some point after and left it there, those kernels beats most of the other not-Yolo kernels that have severals of weeks of work all the way to the deadline, and then the final auto-pick algorithm handled the rest.", "958938": "" } }