![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2109972%2Fe97a5093ae554545493792a8a3692c85%2F34571587690162_.pic.jpg?generation=1587690187389692&alt=media\")
My Team is \"All Faces Are Fake\".\nWe ranked 8th in the previous public leaderboard. \nBut now All of our submissions are shown Error in the end(Invalidated for stage 2 error).\n And our public scores are removed.
\n\nWe have the following questions.\n1. Why we have achieved results in the previous public rankings successfully, but now they have been deleted. I don't think this makes sense because we have scores in the previous public test.\n2. Whether the failed reruns were re-runned again? \n3. Is it possible for us to get our private test logs? It is important.
\n\n*Really really want to figure all these out, since we have spent a lot of time and achieved the 8th place in the public leaderboard. We hope our efforts will not be in vain. *
\n\n
My Team is \"All Faces Are Fake\".\nWe ranked 8th in the previous public leaderboard. \nBut now All of our submissions are shown Error in the end(Invalidated for stage 2 error).\n And our public scores are removed.
\n\nWe have the following questions.\n1. Why we have achieved results in the previous public rankings successfully, but now they have been deleted. I don't think this makes sense because we have scores in the previous public test.\n2. Whether the failed reruns were re-runned again? \n3. Is it possible for us to get our private test logs? It is important.
\n\n*Really really want to figure all these out, since we have spent a lot of time and achieved the 8th place in the public leaderboard. We hope our efforts will not be in vain. *
\n\n
That's very odd. Does it take you to the proper version of your code when you click on the kernel link?
\n\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F1035002%2Fe4acc52974bb749db8aa3114ca0484e2%2FScreen%20Shot%202020-04-23%20at%206.02.36%20PM.png?generation=1587690275964006&alt=media\")
Here's what mine looks like for reference. It shows the public and private scores both went through.
", "rawMarkdown": "That's very odd. Does it take you to the proper version of your code when you click on the kernel link?\n\n\n\nHere's what mine looks like for reference. It shows the public and private scores both went through.", "votes": null }, { "id": "818564", "postDate": "04/24/2020 01:12:48", "content": "The versions of my final submissions are right. I just checked it again.
", "rawMarkdown": "The versions of my final submissions are right. I just checked it again.", "votes": null }, { "id": "818565", "postDate": "04/24/2020 01:13:11", "content": "Same scenario happened to our team, except we only ranked 100+ on the public LB. Really want to figure those things out.
", "rawMarkdown": "Same scenario happened to our team, except we only ranked 100+ on the public LB. Really want to figure those things out.", "votes": null }, { "id": "818587", "postDate": "04/24/2020 01:34:51", "content": "According to Kaggle:
\n\n\n\n\nThere was a ~6.5% failure rate on re-runs, due mostly to incomplete submissions (submissions which failed to create predictions for all samples)
\n
Did you have a failsafe in place to fill in missing predictions for failure cases?
", "rawMarkdown": "According to Kaggle: \n>There was a ~6.5% failure rate on re-runs, due mostly to incomplete submissions (submissions which failed to create predictions for all samples)\n\nDid you have a failsafe in place to fill in missing predictions for failure cases?", "votes": null }, { "id": "818589", "postDate": "04/24/2020 01:40:32", "content": "Our notebook is modified from Human Analog's Inference Demo Notebook.\nThe code handles failures by “try-except” way.
", "rawMarkdown": "Our notebook is modified from Human Analog's Inference Demo Notebook.\nThe code handles failures by “try-except” way.", "votes": null }, { "id": "818609", "postDate": "04/24/2020 02:12:28", "content": "Same, this competition is the worst one I participated.👎 👎 👎 In the beginning, the system is not stable and we need to cost one more time submits to get the final result for some notebooks. Now, all the notebooks went wrong again. I'm quite speechless about the competence of workers from Facebook.
", "rawMarkdown": "Same, this competition is the worst one I participated.👎 👎 👎 In the beginning, the system is not stable and we need to cost one more time submits to get the final result for some notebooks. Now, all the notebooks went wrong again. I'm quite speechless about the competence of workers from Facebook.", "votes": null }, { "id": "818612", "postDate": "04/24/2020 02:16:30", "content": "Maybe they are not done yet. Did anyone hear any announcements?
", "rawMarkdown": "Maybe they are not done yet. Did anyone hear any announcements?", "votes": null }, { "id": "818617", "postDate": "04/24/2020 02:25:04", "content": "Hello, I have sent an email to one of the hosters, and according to his/her meaning, the rerun process should be over.
", "rawMarkdown": "Hello, I have sent an email to one of the hosters, and according to his/her meaning, the rerun process should be over.", "votes": null }, { "id": "818620", "postDate": "04/24/2020 02:27:23", "content": "This looks like a wide problem. I hope we could hear back from Kaggle bout or provide some feedback. I think many people worked very hard on this competition. I got the same issue in one of my submissions, and it's showing both private and public scores in the notebook, which means that it ran successfully.
", "rawMarkdown": "This looks like a wide problem. I hope we could hear back from Kaggle bout or provide some feedback. I think many people worked very hard on this competition. I got the same issue in one of my submissions, and it's showing both private and public scores in the notebook, which means that it ran successfully.", "votes": null }, { "id": "819011", "postDate": "04/24/2020 09:22:42", "content": "\n\n\nand it's showing both private and public scores in the notebook,
\n
Are they identical?
", "rawMarkdown": "> and it's showing both private and public scores in the notebook,\n\nAre they identical?", "votes": null }, { "id": "819052", "postDate": "04/24/2020 10:11:46", "content": "yes
", "rawMarkdown": "yes", "votes": null }, { "id": "819098", "postDate": "04/24/2020 10:56:53", "content": "I think that means the notebook failed and didn't get a private score, so it shows the public score under \"private\" too (which is incorrect, obviously).
", "rawMarkdown": "I think that means the notebook failed and didn't get a private score, so it shows the public score under \"private\" too (which is incorrect, obviously).", "votes": null }, { "id": "819129", "postDate": "04/24/2020 11:27:03", "content": "I have the same problem. We need explanations! No sharing, Public Success. I have used public baseline for submissions:
\n\nhttps://www.kaggle.com/humananalog/inference-demo
\n\nWhy this kernel FAILED?
", "rawMarkdown": "I have the same problem. We need explanations! No sharing, Public Success. I have used public baseline for submissions:\n\nhttps://www.kaggle.com/humananalog/inference-demo\n\nWhy this kernel FAILED?", "votes": null }, { "id": "819570", "postDate": "04/24/2020 17:15:17", "content": "Maybe, but the public score data , is different than the private score. They look the same next to each in the notebook, other but different in private/public leaderboard. Thanks @humananalog always helpful :) . I think manny people are reporting this error, so we will see what happens. I am happy things worked well for you, you helped everyone in the team
", "rawMarkdown": "Maybe, but the public score data , is different than the private score. They look the same next to each in the notebook, other but different in private/public leaderboard. Thanks @humananalog always helpful :) . I think manny people are reporting this error, so we will see what happens. I am happy things worked well for you, you helped everyone in the team", "votes": null }, { "id": "824194", "postDate": "04/28/2020 08:06:28", "content": "🙉Really want to verify whether my approach is robust. (As there is no late submission)\n😄Will be grateful if exists ways to test my model scores on the private dataset.\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2109972%2F6b79feaebbede8228115a54f2189418e%2F671588226285_.pic.jpg?generation=1589041799568316&alt=media\")
Give up.
", "rawMarkdown": "Give up.", "votes": null }, { "id": "884311", "postDate": "06/13/2020 09:25:59", "content": "It looks like the game is over and some discussion boards have been deleted.
\n\nAlthough two months have passed, and I always try to persuade myself not to think of this competition, I still want to know how my pubilc 8th place solution will perform on the private test set.
\n\nWhether will the private test set open to us ?
", "rawMarkdown": "It looks like the game is over and some discussion boards have been deleted.\n\nAlthough two months have passed, and I always try to persuade myself not to think of this competition, I still want to know how my pubilc 8th place solution will perform on the private test set.\n\nWhether will the private test set open to us ?", "votes": null }, { "id": "884339", "postDate": "06/13/2020 09:44:40", "content": "@wengzejia Looks like this, mate. It is the worst organization of competition. Kaggle Team close eyes and ignore all my requests too. I hope Kaggle Community never meet this incident. Dont give up the competitions in the future, you have very good results! 👍
", "rawMarkdown": "wengzejia Looks like this, mate. It is the worst organization of competition. Kaggle Team [close eyes](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/145982) and ignore all my requests too. I hope Kaggle Community never meet this incident. Dont give up the competitions in the future, you have very good results! 👍", "votes": null }, { "id": "1237565", "postDate": "03/14/2021 09:08:43", "content": "您好，可以分享一下DFDC数据集吗？
", "rawMarkdown": "您好，可以分享一下DFDC数据集吗？", "votes": null } ], "comments": [ { "id": 1237565, "author_name": "fanlang", "author_url": "", "post_date": "03/14/2021 09:08:43", "content": "您好，可以分享一下DFDC数据集吗？
", "votes": null, "replies": [] }, { "id": 818552, "author_name": "ryches", "author_url": "", "post_date": "04/24/2020 01:04:54", "content": "That's very odd. Does it take you to the proper version of your code when you click on the kernel link?
\n\n
Here's what mine looks like for reference. It shows the public and private scores both went through.
", "votes": null, "replies": [ { "id": 818564, "author_name": "wengzejia", "author_url": "", "post_date": "04/24/2020 01:12:48", "content": "The versions of my final submissions are right. I just checked it again.
", "votes": null, "replies": [] } ] }, { "id": 818565, "author_name": "lfcpeng17", "author_url": "", "post_date": "04/24/2020 01:13:11", "content": "Same scenario happened to our team, except we only ranked 100+ on the public LB. Really want to figure those things out.
", "votes": null, "replies": [] }, { "id": 818587, "author_name": "vaillant", "author_url": "", "post_date": "04/24/2020 01:34:51", "content": "According to Kaggle:
\n\n\n\n\nThere was a ~6.5% failure rate on re-runs, due mostly to incomplete submissions (submissions which failed to create predictions for all samples)
\n
Did you have a failsafe in place to fill in missing predictions for failure cases?
", "votes": null, "replies": [ { "id": 818589, "author_name": "wengzejia", "author_url": "", "post_date": "04/24/2020 01:40:32", "content": "Our notebook is modified from Human Analog's Inference Demo Notebook.\nThe code handles failures by “try-except” way.
", "votes": null, "replies": [] } ] }, { "id": 818609, "author_name": "fionalxd", "author_url": "", "post_date": "04/24/2020 02:12:28", "content": "Same, this competition is the worst one I participated.👎 👎 👎 In the beginning, the system is not stable and we need to cost one more time submits to get the final result for some notebooks. Now, all the notebooks went wrong again. I'm quite speechless about the competence of workers from Facebook.
", "votes": null, "replies": [] }, { "id": 818612, "author_name": "abualabed", "author_url": "", "post_date": "04/24/2020 02:16:30", "content": "Maybe they are not done yet. Did anyone hear any announcements?
", "votes": null, "replies": [ { "id": 818617, "author_name": "wengzejia", "author_url": "", "post_date": "04/24/2020 02:25:04", "content": "Hello, I have sent an email to one of the hosters, and according to his/her meaning, the rerun process should be over.
", "votes": null, "replies": [] }, { "id": 818620, "author_name": "abualabed", "author_url": "", "post_date": "04/24/2020 02:27:23", "content": "This looks like a wide problem. I hope we could hear back from Kaggle bout or provide some feedback. I think many people worked very hard on this competition. I got the same issue in one of my submissions, and it's showing both private and public scores in the notebook, which means that it ran successfully.
", "votes": null, "replies": [] }, { "id": 819011, "author_name": "humananalog", "author_url": "", "post_date": "04/24/2020 09:22:42", "content": "\n\n\nand it's showing both private and public scores in the notebook,
\n
Are they identical?
", "votes": null, "replies": [] }, { "id": 819052, "author_name": "abualabed", "author_url": "", "post_date": "04/24/2020 10:11:46", "content": "yes
", "votes": null, "replies": [] }, { "id": 819098, "author_name": "humananalog", "author_url": "", "post_date": "04/24/2020 10:56:53", "content": "I think that means the notebook failed and didn't get a private score, so it shows the public score under \"private\" too (which is incorrect, obviously).
", "votes": null, "replies": [] }, { "id": 819570, "author_name": "abualabed", "author_url": "", "post_date": "04/24/2020 17:15:17", "content": "Maybe, but the public score data , is different than the private score. They look the same next to each in the notebook, other but different in private/public leaderboard. Thanks @humananalog always helpful :) . I think manny people are reporting this error, so we will see what happens. I am happy things worked well for you, you helped everyone in the team
", "votes": null, "replies": [] } ] }, { "id": 819129, "author_name": "shonenkov", "author_url": "", "post_date": "04/24/2020 11:27:03", "content": "I have the same problem. We need explanations! No sharing, Public Success. I have used public baseline for submissions:
\n\nhttps://www.kaggle.com/humananalog/inference-demo
\n\nWhy this kernel FAILED?
", "votes": null, "replies": [] }, { "id": 824194, "author_name": "wengzejia", "author_url": "", "post_date": "04/28/2020 08:06:28", "content": "🙉Really want to verify whether my approach is robust. (As there is no late submission)\n😄Will be grateful if exists ways to test my model scores on the private dataset.\n
Give up.
", "votes": null, "replies": [] }, { "id": 884311, "author_name": "wengzejia", "author_url": "", "post_date": "06/13/2020 09:25:59", "content": "It looks like the game is over and some discussion boards have been deleted.
\n\nAlthough two months have passed, and I always try to persuade myself not to think of this competition, I still want to know how my pubilc 8th place solution will perform on the private test set.
\n\nWhether will the private test set open to us ?
", "votes": null, "replies": [ { "id": 884339, "author_name": "shonenkov", "author_url": "", "post_date": "06/13/2020 09:44:40", "content": "@wengzejia Looks like this, mate. It is the worst organization of competition. Kaggle Team close eyes and ignore all my requests too. I hope Kaggle Community never meet this incident. Dont give up the competitions in the future, you have very good results! 👍
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "818542": "My Team is \"All Faces Are Fake\".\nWe ranked 8th in the previous public leaderboard. \nBut now All of our submissions are shown Error in the end(Invalidated for stage 2 error).\n And our public scores are removed.\n\nWe have the following questions.\n1. Why we have achieved results in the previous public rankings successfully, but now they have been deleted. I don't think this makes sense because we have scores in the previous public test.\n2. Whether the failed reruns were re-runned again? \n3. Is it possible for us to get our private test logs? It is important. \n\n**Really really want to figure all these out, since we have spent a lot of time and achieved the 8th place in the public leaderboard. We hope our efforts will not be in vain. **\n\n", "818552": "That's very odd. Does it take you to the proper version of your code when you click on the kernel link?\n\n\n\nHere's what mine looks like for reference. It shows the public and private scores both went through.", "818564": "The versions of my final submissions are right. I just checked it again.", "818565": "Same scenario happened to our team, except we only ranked 100+ on the public LB. Really want to figure those things out.", "818587": "According to Kaggle: \n>There was a ~6.5% failure rate on re-runs, due mostly to incomplete submissions (submissions which failed to create predictions for all samples)\n\nDid you have a failsafe in place to fill in missing predictions for failure cases?", "818589": "Our notebook is modified from Human Analog's Inference Demo Notebook.\nThe code handles failures by “try-except” way.", "818609": "Same, this competition is the worst one I participated.👎 👎 👎 In the beginning, the system is not stable and we need to cost one more time submits to get the final result for some notebooks. Now, all the notebooks went wrong again. I'm quite speechless about the competence of workers from Facebook.", "818612": "Maybe they are not done yet. Did anyone hear any announcements?", "818617": "Hello, I have sent an email to one of the hosters, and according to his/her meaning, the rerun process should be over.", "818620": "This looks like a wide problem. I hope we could hear back from Kaggle bout or provide some feedback. I think many people worked very hard on this competition. I got the same issue in one of my submissions, and it's showing both private and public scores in the notebook, which means that it ran successfully.", "819011": "> and it's showing both private and public scores in the notebook,\n\nAre they identical?", "819052": "yes", "819098": "I think that means the notebook failed and didn't get a private score, so it shows the public score under \"private\" too (which is incorrect, obviously).", "819129": "I have the same problem. We need explanations! No sharing, Public Success. I have used public baseline for submissions:\n\nhttps://www.kaggle.com/humananalog/inference-demo\n\nWhy this kernel FAILED?", "819570": "Maybe, but the public score data , is different than the private score. They look the same next to each in the notebook, other but different in private/public leaderboard. Thanks @humananalog always helpful :) . I think manny people are reporting this error, so we will see what happens. I am happy things worked well for you, you helped everyone in the team", "824194": "🙉Really want to verify whether my approach is robust. (As there is no late submission)\n😄Will be grateful if exists ways to test my model scores on the private dataset.\n", "839878": "Give up.", "884311": "It looks like the game is over and some discussion boards have been deleted.\n\nAlthough two months have passed, and I always try to persuade myself not to think of this competition, I still want to know how my pubilc 8th place solution will perform on the private test set.\n\nWhether will the private test set open to us ?", "884339": "wengzejia Looks like this, mate. It is the worst organization of competition. Kaggle Team [close eyes](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/145982) and ignore all my requests too. I hope Kaggle Community never meet this incident. Dont give up the competitions in the future, you have very good results! 👍", "1237565": "您好，可以分享一下DFDC数据集吗？" }, "source": "meta" }