might be a silly question but i am eagerly waiting to know why only 530 teams working for this competition?\ni think this competition deserved much more participants than that,am i wrong?
", "messages": [ { "id": "669773", "postDate": "11/10/2019 12:55:12", "content": "might be a silly question but i am eagerly waiting to know why only 530 teams working for this competition?\ni think this competition deserved much more participants than that,am i wrong?
", "rawMarkdown": "might be a silly question but i am eagerly waiting to know why only 530 teams working for this competition?\ni think this competition deserved much more participants than that,am i wrong?", "votes": null }, { "id": "669860", "postDate": "11/10/2019 15:22:37", "content": "I think the big amount of data, high hardware demands and also that it is a new kind of competition are some reasons.
", "rawMarkdown": "I think the big amount of data, high hardware demands and also that it is a new kind of competition are some reasons.", "votes": null }, { "id": "669927", "postDate": "11/10/2019 17:41:01", "content": "rsna dataset size 181 gb but i saw a lot of competitors there
", "rawMarkdown": "rsna dataset size 181 gb but i saw a lot of competitors there", "votes": null }, { "id": "669950", "postDate": "11/10/2019 18:40:56", "content": "I believe that the sheer complexity of a problem and the fact, that this topic is quite new, compared to 2D classification, are among the main reasons. I joined on the last day, and I'm still very confused. There're so many things, that are completely new to me
", "rawMarkdown": "I believe that the sheer complexity of a problem and the fact, that this topic is quite new, compared to 2D classification, are among the main reasons. I joined on the last day, and I'm still very confused. There're so many things, that are completely new to me", "votes": null }, { "id": "670532", "postDate": "11/11/2019 14:56:25", "content": "I would say 100 teams working, and 50-100 more just running public kernels. \nAll others submitted a week ago or even more.
\n\nReasons\n- Kaggle GPU limit\n- A lot of data\n- Not clear competition
", "rawMarkdown": "I would say 100 teams working, and 50-100 more just running public kernels. \nAll others submitted a week ago or even more. \n\nReasons\n- Kaggle GPU limit\n- A lot of data\n- Not clear competition", "votes": null }, { "id": "670536", "postDate": "11/11/2019 14:59:01", "content": "agree,,more grandmaster working in this competition means we would get much better public kernels and it would help us learn more about this competition i think
", "rawMarkdown": "agree,,more grandmaster working in this competition means we would get much better public kernels and it would help us learn more about this competition i think", "votes": null }, { "id": "671610", "postDate": "11/13/2019 00:18:41", "content": "Because it's difficult?
", "rawMarkdown": "Because it's difficult?", "votes": null }, { "id": "671612", "postDate": "11/13/2019 00:23:02", "content": "I'd say hardware requirements for Open Images were a lot higher. I blame an unknown field and the lack of well-known practices. If we're talking about semantic segmentation or image classification, there are plenty of ready pipelines at GitHub, lots of articles, etc.
", "rawMarkdown": "I'd say hardware requirements for Open Images were a lot higher. I blame an unknown field and the lack of well-known practices. If we're talking about semantic segmentation or image classification, there are plenty of ready pipelines at GitHub, lots of articles, etc.", "votes": null }, { "id": "671615", "postDate": "11/13/2019 00:26:39", "content": "object detection with point cloud is a pre-mature area in the phase of research.\nI hope that this competition will attract more people into lidar processing in their fields of research/works.. :)
", "rawMarkdown": "object detection with point cloud is a pre-mature area in the phase of research.\nI hope that this competition will attract more people into lidar processing in their fields of research/works.. :)", "votes": null }, { "id": "671640", "postDate": "11/13/2019 01:23:36", "content": "Yes, maybe. I used second.pytorch which after going through the hassle of writing the importer etc. worked great. Anyways i am happy, that there haven't been more grandmasters :-)
", "rawMarkdown": "Yes, maybe. I used second.pytorch which after going through the hassle of writing the importer etc. worked great. Anyways i am happy, that there haven't been more grandmasters :-)", "votes": null }, { "id": "671647", "postDate": "11/13/2019 01:45:33", "content": "I tried second but didn't work for me. The code is too complicated for me. May you share your experience?
", "rawMarkdown": "I tried second but didn't work for me. The code is too complicated for me. May you share your experience?", "votes": null }, { "id": "671666", "postDate": "11/13/2019 02:36:50", "content": "classification task is easier to start : )
", "rawMarkdown": "classification task is easier to start : )", "votes": null }, { "id": "672059", "postDate": "11/13/2019 13:27:14", "content": "Bob, i described it in the \"What model did you use?\" thread.
", "rawMarkdown": "Bob, i described it in the \"What model did you use?\" thread.", "votes": null } ], "comments": [ { "id": 669860, "author_name": "marekwyborski", "author_url": "", "post_date": "11/10/2019 15:22:37", "content": "I think the big amount of data, high hardware demands and also that it is a new kind of competition are some reasons.
", "votes": null, "replies": [ { "id": 669927, "author_name": "mobassir", "author_url": "", "post_date": "11/10/2019 17:41:01", "content": "rsna dataset size 181 gb but i saw a lot of competitors there
", "votes": null, "replies": [] }, { "id": 669950, "author_name": "cateek", "author_url": "", "post_date": "11/10/2019 18:40:56", "content": "I believe that the sheer complexity of a problem and the fact, that this topic is quite new, compared to 2D classification, are among the main reasons. I joined on the last day, and I'm still very confused. There're so many things, that are completely new to me
", "votes": null, "replies": [] }, { "id": 671612, "author_name": "artyomp", "author_url": "", "post_date": "11/13/2019 00:23:02", "content": "I'd say hardware requirements for Open Images were a lot higher. I blame an unknown field and the lack of well-known practices. If we're talking about semantic segmentation or image classification, there are plenty of ready pipelines at GitHub, lots of articles, etc.
", "votes": null, "replies": [] }, { "id": 671615, "author_name": "kyoshioka47", "author_url": "", "post_date": "11/13/2019 00:26:39", "content": "object detection with point cloud is a pre-mature area in the phase of research.\nI hope that this competition will attract more people into lidar processing in their fields of research/works.. :)
", "votes": null, "replies": [] }, { "id": 671640, "author_name": "marekwyborski", "author_url": "", "post_date": "11/13/2019 01:23:36", "content": "Yes, maybe. I used second.pytorch which after going through the hassle of writing the importer etc. worked great. Anyways i am happy, that there haven't been more grandmasters :-)
", "votes": null, "replies": [] }, { "id": 671647, "author_name": "usherbob", "author_url": "", "post_date": "11/13/2019 01:45:33", "content": "I tried second but didn't work for me. The code is too complicated for me. May you share your experience?
", "votes": null, "replies": [] }, { "id": 671666, "author_name": "yiheng", "author_url": "", "post_date": "11/13/2019 02:36:50", "content": "classification task is easier to start : )
", "votes": null, "replies": [] }, { "id": 672059, "author_name": "marekwyborski", "author_url": "", "post_date": "11/13/2019 13:27:14", "content": "Bob, i described it in the \"What model did you use?\" thread.
", "votes": null, "replies": [] } ] }, { "id": 670532, "author_name": "muhakabartay", "author_url": "", "post_date": "11/11/2019 14:56:25", "content": "I would say 100 teams working, and 50-100 more just running public kernels. \nAll others submitted a week ago or even more.
\n\nReasons\n- Kaggle GPU limit\n- A lot of data\n- Not clear competition
", "votes": null, "replies": [ { "id": 670536, "author_name": "mobassir", "author_url": "", "post_date": "11/11/2019 14:59:01", "content": "agree,,more grandmaster working in this competition means we would get much better public kernels and it would help us learn more about this competition i think
", "votes": null, "replies": [] } ] }, { "id": 671610, "author_name": "artyomp", "author_url": "", "post_date": "11/13/2019 00:18:41", "content": "Because it's difficult?
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "669773": "might be a silly question but i am eagerly waiting to know why only 530 teams working for this competition?\ni think this competition deserved much more participants than that,am i wrong?", "669860": "I think the big amount of data, high hardware demands and also that it is a new kind of competition are some reasons.", "669927": "rsna dataset size 181 gb but i saw a lot of competitors there", "669950": "I believe that the sheer complexity of a problem and the fact, that this topic is quite new, compared to 2D classification, are among the main reasons. I joined on the last day, and I'm still very confused. There're so many things, that are completely new to me", "670532": "I would say 100 teams working, and 50-100 more just running public kernels. \nAll others submitted a week ago or even more. \n\nReasons\n- Kaggle GPU limit\n- A lot of data\n- Not clear competition", "670536": "agree,,more grandmaster working in this competition means we would get much better public kernels and it would help us learn more about this competition i think", "671610": "Because it's difficult?", "671612": "I'd say hardware requirements for Open Images were a lot higher. I blame an unknown field and the lack of well-known practices. If we're talking about semantic segmentation or image classification, there are plenty of ready pipelines at GitHub, lots of articles, etc.", "671615": "object detection with point cloud is a pre-mature area in the phase of research.\nI hope that this competition will attract more people into lidar processing in their fields of research/works.. :)", "671640": "Yes, maybe. I used second.pytorch which after going through the hassle of writing the importer etc. worked great. Anyways i am happy, that there haven't been more grandmasters :-)", "671647": "I tried second but didn't work for me. The code is too complicated for me. May you share your experience?", "671666": "classification task is easier to start : )", "672059": "Bob, i described it in the \"What model did you use?\" thread." }, "source": "meta" }