I just wanted to ask when the new dataset will be made public. Will it take some days or weeks ?
", "messages": [ { "id": 1356283, "postDate": "2021-06-18T21:34:39.840Z", "content": "I'll provide an estimate of timing on Monday.
", "rawMarkdown": "I'll provide an estimate of timing on Monday.", "votes": 11, "replies": [ { "id": 1356303, "postDate": "2021-06-18T22:07:32.257Z", "content": "I'm sure this is a tough situation for you. We all hope you'll find a good way out of it. And we can still work, it is not as if we had to wait for you to do something useful.
", "rawMarkdown": "I'm sure this is a tough situation for you. We all hope you'll find a good way out of it. And we can still work, it is not as if we had to wait for you to do something useful.", "votes": 7 }, { "id": 1356327, "postDate": "2021-06-18T23:08:47.450Z", "content": "Thanks for letting us now. This competition is great and it keeps being so.
", "rawMarkdown": "Thanks for letting us now. This competition is great and it keeps being so. ", "votes": 1 }, { "id": 1356406, "postDate": "2021-06-19T02:17:30.993Z", "content": "@inversion Thank you for your quick response and hard work. I will carefully check for other leaks when new data is released.
", "rawMarkdown": "@inversion Thank you for your quick response and hard work. I will carefully check for other leaks when new data is released.", "votes": 4 }, { "id": 1356427, "postDate": "2021-06-19T02:54:06.767Z", "content": "@inv136356 Thank You for your reply 😃
", "rawMarkdown": "@inv136356 Thank You for your reply 😃", "votes": 1 }, { "id": 1356814, "postDate": "2021-06-19T08:21:48.207Z", "content": "@inversion I am pretty sure this is unexpected and a tall order for you and your team to get the data ready on such short notice. I am constantly fascinated by the feats Kaggle pulls off, so I know you people can manage this as well. No pressure, please take your time and be through with the checks so that we address all known issues. Don't lose your sleep or weekend over it, Its going to be just fine…
", "rawMarkdown": "@inversion I am pretty sure this is unexpected and a tall order for you and your team to get the data ready on such short notice. I am constantly fascinated by the feats Kaggle pulls off, so I know you people can manage this as well. No pressure, please take your time and be through with the checks so that we address all known issues. Don't lose your sleep or weekend over it, Its going to be just fine...", "votes": 5 }, { "id": 1360565, "postDate": "2021-06-22T07:16:28.997Z", "content": "@inversion any updates?
", "rawMarkdown": "@inversion any updates?" }, { "id": 1360582, "postDate": "2021-06-22T07:36:24.367Z", "content": "Still working hard on the new version. I'm reluctant to provide any concrete timeline at this moment as we want to make sure we plugged all the leaks we could find.
", "rawMarkdown": "Still working hard on the new version. I'm reluctant to provide any concrete timeline at this moment as we want to make sure we plugged all the leaks we could find.", "votes": 24 }, { "id": 1360640, "postDate": "2021-06-22T08:43:59.587Z", "content": "Thank you for your work! And it seems better not to rush and check everything several times.
", "rawMarkdown": "Thank you for your work! And it seems better not to rush and check everything several times.", "votes": 5 } ] }, { "id": 1355171, "postDate": "2021-06-18T06:24:50.573Z", "content": "I just wanted to ask when the new dataset will be made public. Will it take some days or weeks ?
", "rawMarkdown": "I just wanted to ask when the new dataset will be made public. Will it take some days or weeks ?", "votes": 7 }, { "id": 1355479, "postDate": "2021-06-18T09:47:52.363Z", "content": "It is better if they take time to double check every reported issues rather than hurry.
\nYou can use the provided datasets in the meantime. To remove leaks, just ignore meta data and normalize all your snippets when you load them. This way you have data immune from the two reported leaks hopefully.
", "rawMarkdown": "It is better if they take time to double check every reported issues rather than hurry.\n\nYou can use the provided datasets in the meantime. To remove leaks, just ignore meta data and normalize all your snippets when you load them. This way you have data immune from the two reported leaks hopefully.", "votes": 5, "replies": [ { "id": 1355600, "postDate": "2021-06-18T11:58:24.897Z", "content": "What about the downcast truncation mentioned by @sggpls ? Should the data be in int16 to avoid the float issue?
\nhttps://www.kaggle.com/c/seti-breakthrough-listen/discussion/246901
", "rawMarkdown": "What about the downcast truncation mentioned by @sggpls ? Should the data be in int16 to avoid the float issue?\n\nhttps://www.kaggle.com/c/seti-breakthrough-listen/discussion/246901", "votes": 2 }, { "id": 1356302, "postDate": "2021-06-18T22:06:29.223Z", "content": "FP16 is not the issue. There is a leak whenever there is a difference between how snippets without messages , and snippets with messages are processed.
", "rawMarkdown": "FP16 is not the issue. There is a leak whenever there is a difference between how snippets without messages , and snippets with messages are processed.", "votes": 3 }, { "id": 1356678, "postDate": "2021-06-19T06:55:42.297Z", "content": "Agreed, do not give pressure on the host. We can totally do experiments without leakage and LB.
", "rawMarkdown": "Agreed, do not give pressure on the host. We can totally do experiments without leakage and LB.", "votes": 1 } ] } ], "comments": [ { "id": 1356283, "author_name": "inversion", "author_url": "", "post_date": "2021-06-18T21:34:39.840000", "content": "I'll provide an estimate of timing on Monday.
", "votes": 11, "replies": [ { "id": 1356303, "author_name": "CPMP", "author_url": "", "post_date": "2021-06-18T22:07:32.257000", "content": "I'm sure this is a tough situation for you. We all hope you'll find a good way out of it. And we can still work, it is not as if we had to wait for you to do something useful.
", "votes": 7, "replies": [] }, { "id": 1356327, "author_name": "FelipeKitamura, MD, PhD", "author_url": "", "post_date": "2021-06-18T23:08:47.450000", "content": "Thanks for letting us now. This competition is great and it keeps being so.
", "votes": 1, "replies": [] }, { "id": 1356406, "author_name": "nyanp", "author_url": "", "post_date": "2021-06-19T02:17:30.993000", "content": "@inversion Thank you for your quick response and hard work. I will carefully check for other leaks when new data is released.
", "votes": 4, "replies": [] }, { "id": 1356427, "author_name": "Mithil Salunkhe", "author_url": "", "post_date": "2021-06-19T02:54:06.767000", "content": "@inv136356 Thank You for your reply 😃
", "votes": 1, "replies": [] }, { "id": 1356814, "author_name": "Manav", "author_url": "", "post_date": "2021-06-19T08:21:48.207000", "content": "@inversion I am pretty sure this is unexpected and a tall order for you and your team to get the data ready on such short notice. I am constantly fascinated by the feats Kaggle pulls off, so I know you people can manage this as well. No pressure, please take your time and be through with the checks so that we address all known issues. Don't lose your sleep or weekend over it, Its going to be just fine…
", "votes": 5, "replies": [] }, { "id": 1360565, "author_name": "Psi", "author_url": "", "post_date": "2021-06-22T07:16:28.997000", "content": "@inversion any updates?
", "votes": 0, "replies": [] }, { "id": 1360582, "author_name": "Yuhong Chen", "author_url": "", "post_date": "2021-06-22T07:36:24.367000", "content": "Still working hard on the new version. I'm reluctant to provide any concrete timeline at this moment as we want to make sure we plugged all the leaks we could find.
", "votes": 24, "replies": [] }, { "id": 1360640, "author_name": "Sergey Bryansky", "author_url": "", "post_date": "2021-06-22T08:43:59.587000", "content": "Thank you for your work! And it seems better not to rush and check everything several times.
", "votes": 5, "replies": [] } ] }, { "id": 1355479, "author_name": "CPMP", "author_url": "", "post_date": "2021-06-18T09:47:52.363000", "content": "It is better if they take time to double check every reported issues rather than hurry.
\nYou can use the provided datasets in the meantime. To remove leaks, just ignore meta data and normalize all your snippets when you load them. This way you have data immune from the two reported leaks hopefully.
", "votes": 5, "replies": [ { "id": 1355600, "author_name": "FelipeKitamura, MD, PhD", "author_url": "", "post_date": "2021-06-18T11:58:24.897000", "content": "What about the downcast truncation mentioned by @sggpls ? Should the data be in int16 to avoid the float issue?
\nhttps://www.kaggle.com/c/seti-breakthrough-listen/discussion/246901
", "votes": 2, "replies": [] }, { "id": 1356302, "author_name": "CPMP", "author_url": "", "post_date": "2021-06-18T22:06:29.223000", "content": "FP16 is not the issue. There is a leak whenever there is a difference between how snippets without messages , and snippets with messages are processed.
", "votes": 3, "replies": [] }, { "id": 1356678, "author_name": "Chenglu", "author_url": "", "post_date": "2021-06-19T06:55:42.297000", "content": "Agreed, do not give pressure on the host. We can totally do experiments without leakage and LB.
", "votes": 1, "replies": [] } ] } ], "raw_markdown_by_id": { "1356283": "I'll provide an estimate of timing on Monday.", "1355171": "I just wanted to ask when the new dataset will be made public. Will it take some days or weeks ?", "1355479": "It is better if they take time to double check every reported issues rather than hurry.\n\nYou can use the provided datasets in the meantime. To remove leaks, just ignore meta data and normalize all your snippets when you load them. This way you have data immune from the two reported leaks hopefully." } }