\n
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3613976%2F513b6ad67fcfd3de5848ed8a4cbb7c16%2FSelection_317.png?generation=1597869629308451&alt=media\")
\nOld scores
\n
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3613976%2F4a90e4183a8210222cfdc243af4b6dd6%2F012946e9-47e0-4267-9445-6e0913e018b7.jpeg?generation=1597871158557972&alt=media\")
My old private scores were not updated, and I was mistakenly placed #1 in the competition. I did not choose 2 of my final submissions.
\nNew scores
\n
\nOld scores
\n
As you can see from the screenshots, my scores for final submissions are exactly the same as they were for the old private set. My assumption (just a speculation) is that the notebooks of the participants who did not choose their final submissions were not rerun on the new test set.
\nThere might be more such cases, so I think the kaggle team should take a day or two to fix the bug, and finally provide fair results …
\nEdit: I was simply removed from the competition … I hope they will bring me back …
", "messages": [ { "id": 978006, "postDate": "2020-08-19T21:06:58.500Z", "content": "My old private scores were not updated, and I was mistakenly placed #1 in the competition. I did not choose 2 of my final submissions.
\nNew scores
\n
\nOld scores
\n
As you can see from the screenshots, my scores for final submissions are exactly the same as they were for the old private set. My assumption (just a speculation) is that the notebooks of the participants who did not choose their final submissions were not rerun on the new test set.
\nThere might be more such cases, so I think the kaggle team should take a day or two to fix the bug, and finally provide fair results …
\nEdit: I was simply removed from the competition … I hope they will bring me back …
", "rawMarkdown": "My old private scores were not updated, and I was mistakenly placed #1 in the competition. I did not choose 2 of my final submissions. \n\nNew scores\n\nOld scores\n\n\nAs you can see from the screenshots, my scores for final submissions are exactly the same as they were for the old private set. My assumption (just a speculation) is that the notebooks of the participants who did not choose their final submissions were not rerun on the new test set. \n\nThere might be more such cases, so I think the kaggle team should take a day or two to fix the bug, and finally provide fair results ...\n\nEdit: I was simply removed from the competition ... I hope they will bring me back ...", "votes": 33 }, { "id": 978968, "postDate": "2020-08-20T14:41:41.770Z", "content": "This is a serious bug, kaggle team please look into this matter!
", "rawMarkdown": "This is a serious bug, kaggle team please look into this matter!", "votes": 7 }, { "id": 978045, "postDate": "2020-08-19T22:13:00.933Z", "content": "Interesting. I just made a similar post before seeing this one. I thought perhaps the reason was because I had deleted my dataset; but like you, I too did not select final submissions and just let kaggle to its thing. At the point of the competition close, I only had 3 successful submissions anyway.
\nSo either you're right that it has to do with not selecting successful submissions, or perhaps faster-rcnn was ban-hammered because my models that now show up as \"Error\" were also faster-rcnn based. Let's see how it goes =).
", "rawMarkdown": "Interesting. I just made a similar post before seeing this one. I thought perhaps the reason was because I had deleted my dataset; but like you, I too did not select final submissions and just let kaggle to its thing. At the point of the competition close, I only had 3 successful submissions anyway.\n\nSo either you're right that it has to do with not selecting successful submissions, or perhaps faster-rcnn was ban-hammered because my models that now show up as \"Error\" were also faster-rcnn based. Let's see how it goes =).", "votes": 2, "replies": [ { "id": 978053, "postDate": "2020-08-19T22:35:00.507Z", "content": "I was mistakenly placed #1 in the competition due to this bug (not updated scores), and they did not remove me from the competition until I reported this bug to @philculliton . In your case, it might be because you deleted the dataset … The hosts/kaggle team simply reran the notebooks instead of changing the evaluation script, so I guess your notebook failed due to the deleted dataset.
", "rawMarkdown": "I was mistakenly placed #1 in the competition due to this bug (not updated scores), and they did not remove me from the competition until I reported this bug to @philculliton . In your case, it might be because you deleted the dataset ... The hosts/kaggle team simply reran the notebooks instead of changing the evaluation script, so I guess your notebook failed due to the deleted dataset. ", "votes": 10 } ] }, { "id": 1370453, "postDate": "2021-06-30T07:17:59.877Z", "content": "This was very helpful!
", "rawMarkdown": "This was very helpful!" }, { "id": 978952, "postDate": "2020-08-20T14:22:52.733Z", "rawMarkdown": "", "isDeleted": true }, { "id": 978801, "postDate": "2020-08-20T11:58:21.823Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 978968, "author_name": "alan chen", "author_url": "", "post_date": "2020-08-20T14:41:41.770000", "content": "This is a serious bug, kaggle team please look into this matter!
", "votes": 7, "replies": [] }, { "id": 978045, "author_name": "عثمان", "author_url": "", "post_date": "2020-08-19T22:13:00.933000", "content": "Interesting. I just made a similar post before seeing this one. I thought perhaps the reason was because I had deleted my dataset; but like you, I too did not select final submissions and just let kaggle to its thing. At the point of the competition close, I only had 3 successful submissions anyway.
\nSo either you're right that it has to do with not selecting successful submissions, or perhaps faster-rcnn was ban-hammered because my models that now show up as \"Error\" were also faster-rcnn based. Let's see how it goes =).
", "votes": 2, "replies": [ { "id": 978053, "author_name": "Aidyn Ubingazhibov", "author_url": "", "post_date": "2020-08-19T22:35:00.507000", "content": "I was mistakenly placed #1 in the competition due to this bug (not updated scores), and they did not remove me from the competition until I reported this bug to @philculliton . In your case, it might be because you deleted the dataset … The hosts/kaggle team simply reran the notebooks instead of changing the evaluation script, so I guess your notebook failed due to the deleted dataset.
", "votes": 10, "replies": [] } ] }, { "id": 1370453, "author_name": "Justin Chen", "author_url": "", "post_date": "2021-06-30T07:17:59.877000", "content": "This was very helpful!
", "votes": 0, "replies": [] }, { "id": 978952, "author_name": "", "author_url": "", "post_date": "2020-08-20T14:22:52.733000", "content": "", "votes": 0, "replies": [] }, { "id": 978801, "author_name": "", "author_url": "", "post_date": "2020-08-20T11:58:21.823000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "978006": "My old private scores were not updated, and I was mistakenly placed #1 in the competition. I did not choose 2 of my final submissions. \n\nNew scores\n\nOld scores\n\n\nAs you can see from the screenshots, my scores for final submissions are exactly the same as they were for the old private set. My assumption (just a speculation) is that the notebooks of the participants who did not choose their final submissions were not rerun on the new test set. \n\nThere might be more such cases, so I think the kaggle team should take a day or two to fix the bug, and finally provide fair results ...\n\nEdit: I was simply removed from the competition ... I hope they will bring me back ...", "978968": "This is a serious bug, kaggle team please look into this matter!", "978045": "Interesting. I just made a similar post before seeing this one. I thought perhaps the reason was because I had deleted my dataset; but like you, I too did not select final submissions and just let kaggle to its thing. At the point of the competition close, I only had 3 successful submissions anyway.\n\nSo either you're right that it has to do with not selecting successful submissions, or perhaps faster-rcnn was ban-hammered because my models that now show up as \"Error\" were also faster-rcnn based. Let's see how it goes =).", "1370453": "This was very helpful!", "978952": "", "978801": "" } }