We try to form a team now with @yassinealouini \n@htopper @stanislavblinov would be might great to slightly increase the size of our team to be 4-5 in total (I pretty sure that more than 4-5 people in a team lead to degradation of efficiency). \nLet's see the opinion of @yassinealouini \nI stand with the more less same idea as @stanislavblinov, but to bring another method would be great since I don't think so that with LGBM or whatsoever we can even touch the LB Top100 ))). But neural networks power the top10 LB as we know.
", "rawMarkdown": "We try to form a team now with @yassinealouini \n@htopper @stanislavblinov would be might great to slightly increase the size of our team to be 4-5 in total (I pretty sure that more than 4-5 people in a team lead to degradation of efficiency). \nLet's see the opinion of @yassinealouini \nI stand with the more less same idea as @stanislavblinov, but to bring another method would be great since I don't think so that with LGBM or whatsoever we can even touch the LB Top100 ))). But neural networks power the top10 LB as we know.", "votes": 1, "replies": [ { "id": 585042, "postDate": "2019-07-26T20:45:43.307Z", "content": "Sounds like a good idea. Let us discuss via email if you want. Mine is yassinealouini@outlook.com.
", "rawMarkdown": "Sounds like a good idea. Let us discuss via email if you want. Mine is yassinealouini@outlook.com.", "votes": 1 }, { "id": 585066, "postDate": "2019-07-26T21:46:21.200Z", "content": "Add me to the thread if you start.\nOrganokov.m@gmail.com
", "rawMarkdown": "Add me to the thread if you start.\nOrganokov.m@gmail.com ", "votes": 1 }, { "id": 585109, "postDate": "2019-07-27T01:21:58.727Z", "content": "Hey, it's great to so see you guys grouping up
\n\nBest of luck on this competition!
", "rawMarkdown": "Hey, it's great to so see you guys grouping up\n\nBest of luck on this competition!", "votes": 1 }, { "id": 598986, "postDate": "2019-08-14T10:28:51.090Z", "content": "Hi guys! @muhakabartay @yassinealouini \nI'm kinda back into business (a little later than expected tho). Is your invitation still up?
", "rawMarkdown": "Hi guys! @muhakabartay @yassinealouini \nI'm kinda back into business (a little later than expected tho). Is your invitation still up?", "votes": 1 }, { "id": 599064, "postDate": "2019-08-14T12:56:26.410Z", "content": "Hi Stanislav @stanislavblinov
\n\nSome things changed but not vastly. I just a few days before I discussed with our team @yassinealouini and @harshit92 that we need someone with the model as you have. Because it can boost our score if you really reached -1.7 only via the usage of LGBM model. @yassinealouini @harshit92 what do you think of that if Stanislav will join us and two great and different models can be brought together?
\n\nP.S. @yassinealouini I told you (don't want to tell it obvious here how) that such exact -1.7 in LGBM can boost us. And of course, can boost Stanislav. but both teams able to fight for bronze I think (our silver is under big questions and with Stanislav's model we can fight for silver with higher confidence than now).
", "rawMarkdown": "Hi Stanislav @stanislavblinov \n\nSome things changed but not vastly. I just a few days before I discussed with our team @yassinealouini and @harshit92 that we need someone with the model as you have. Because it can boost our score if you really reached -1.7 only via the usage of LGBM model. @yassinealouini @harshit92 what do you think of that if Stanislav will join us and two great and different models can be brought together? \n\nP.S. @yassinealouini I told you (don't want to tell it obvious here how) that such exact -1.7 in LGBM can boost us. And of course, can boost Stanislav. but both teams able to fight for bronze I think (our silver is under big questions and with Stanislav's model we can fight for silver with higher confidence than now). ", "votes": 3 }, { "id": 599083, "postDate": "2019-08-14T13:24:09.480Z", "content": "@muhakabartay \nMy LGB-and-nothing-else model has CV -1.78 and LB -1.73.
\n\n\n\n\nwe can fight for silver
\n
Nah, green zone only :)
", "rawMarkdown": "@muhakabartay \nMy LGB-and-nothing-else model has CV -1.78 and LB -1.73.\n> we can fight for silver\n\nNah, green zone only :)", "votes": 1 }, { "id": 599277, "postDate": "2019-08-14T18:42:21.767Z", "content": "@stanislavblinov even better ))))))))))
\nLet's wait response of the other my teammates.
I have sent you a merge request. Hopefully our collaboration is fruitful.
", "rawMarkdown": "I have sent you a merge request. Hopefully our collaboration is fruitful. ", "votes": 2 }, { "id": 599639, "postDate": "2019-08-15T07:59:19.027Z", "content": "Accepted, thanks. \nAlso, I sent you an email
", "rawMarkdown": "Accepted, thanks. \nAlso, I sent you an email", "votes": 2 } ] }, { "id": 581370, "postDate": "2019-07-21T20:45:10.980Z", "content": "Hi, I'm interested!\nAbout me:\nI am quite new to kaggle though already earned a couple of medals, and not very experienced in data science in general (did my very first print('hello, world') ~7 months ago) and therefore would really love to have an experienced teammate / s.\nAbout my model:\nMy model is GPU-accelerated LGBM, mean CV -1.64, LB -1.57.\nImportant: I'm on vacation at the moment without access to my computer, therefore I won't be able to participate till 31st of July. I am ready to put some other things aside and make more spare time to spend on this comp in August (discussable 😉 )
\n\nAnyhow, if you're interested, you can contact me here on kaggle or via e-mail blinberg@mail.ru.
", "rawMarkdown": "Hi, I'm interested!\n**About me:**\nI am quite new to kaggle though already earned a couple of medals, and not very experienced in data science in general (did my very first print('hello, world') ~7 months ago) and therefore would really love to have an experienced teammate / s.\n**About my model:**\nMy model is GPU-accelerated LGBM, mean CV -1.64, LB -1.57.\n**Important:** I'm on vacation at the moment without access to my computer, therefore I won't be able to participate till 31st of July. I am ready to put some other things aside and make more spare time to spend on this comp in August (discussable 😉 )\n\nAnyhow, if you're interested, you can contact me here on kaggle or via e-mail blinberg@mail.ru.", "votes": 1, "replies": [ { "id": 581696, "postDate": "2019-07-22T09:14:26.877Z", "content": "I wish you have a great vacation and I hope you spend it happily!
", "rawMarkdown": "I wish you have a great vacation and I hope you spend it happily!", "votes": 4 } ] }, { "id": 581118, "postDate": "2019-07-21T12:54:32.120Z", "content": "I can probably add some diversity to your models. I'm exclusively using graph cnns. My best LB score is from a single model trained on all coupling types.
", "rawMarkdown": "I can probably add some diversity to your models. I'm exclusively using graph cnns. My best LB score is from a single model trained on all coupling types.", "votes": 1, "replies": [ { "id": 581140, "postDate": "2019-07-21T13:56:25.470Z", "rawMarkdown": "" } ] }, { "id": 584938, "postDate": "2019-07-26T17:40:30.480Z", "rawMarkdown": "", "isDeleted": true, "replies": [ { "id": 585111, "postDate": "2019-07-27T01:25:54.873Z", "content": "We are using lightgbm and 2 different graph nn, the high score is from a combination of those \n(no single model above -1.9)
", "rawMarkdown": "We are using lightgbm and 2 different graph nn, the high score is from a combination of those \n(no single model above -1.9)", "votes": 1 }, { "id": 588482, "postDate": "2019-07-30T16:19:40.167Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 581106, "postDate": "2019-07-21T12:38:44.930Z", "rawMarkdown": "", "votes": 2, "isDeleted": true, "replies": [ { "id": 581675, "postDate": "2019-07-22T08:43:04.550Z", "content": "here
\n\nP.S: I am a fan of Robin VanPersi.
", "rawMarkdown": "here\n\nP.S: I am a fan of Robin VanPersi.", "votes": 1 } ] }, { "id": 580783, "postDate": "2019-07-20T18:47:00.840Z", "rawMarkdown": "", "isDeleted": true, "replies": [ { "id": 581035, "postDate": "2019-07-21T09:38:17.490Z", "content": "We are looking for teammates with a better score than highest public kernel - 0.15
", "rawMarkdown": "We are looking for teammates with a better score than *highest public kernel - 0.15*" }, { "id": 584377, "postDate": "2019-07-25T20:56:55.730Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 588481, "postDate": "2019-07-30T16:18:56.653Z", "content": "@doctorgg might join our team?
\n\n@yassinealouini someone from top80 might bring good advancements. What do you think?
", "rawMarkdown": "@doctorgg might join our team? \n\n@yassinealouini someone from top80 might bring good advancements. What do you think? ", "votes": 1 }, { "id": 588535, "postDate": "2019-07-30T18:14:28.397Z", "content": "Yes for sure, let me know if it is fine with you @doctorgg and I will send an invitation. :)
", "rawMarkdown": "Yes for sure, let me know if it is fine with you @doctorgg and I will send an invitation. :)" } ] } ], "comments": [ { "id": 584944, "author_name": "Mukharbek Organokov", "author_url": "", "post_date": "2019-07-26T17:47:51.683000", "content": "We try to form a team now with @yassinealouini \n@htopper @stanislavblinov would be might great to slightly increase the size of our team to be 4-5 in total (I pretty sure that more than 4-5 people in a team lead to degradation of efficiency). \nLet's see the opinion of @yassinealouini \nI stand with the more less same idea as @stanislavblinov, but to bring another method would be great since I don't think so that with LGBM or whatsoever we can even touch the LB Top100 ))). But neural networks power the top10 LB as we know.
", "votes": 1, "replies": [ { "id": 585042, "author_name": "Yassine Alouini", "author_url": "", "post_date": "2019-07-26T20:45:43.307000", "content": "Sounds like a good idea. Let us discuss via email if you want. Mine is yassinealouini@outlook.com.
", "votes": 1, "replies": [] }, { "id": 585066, "author_name": "Mukharbek Organokov", "author_url": "", "post_date": "2019-07-26T21:46:21.200000", "content": "Add me to the thread if you start.\nOrganokov.m@gmail.com
", "votes": 1, "replies": [] }, { "id": 585109, "author_name": "Cozy Doomer", "author_url": "", "post_date": "2019-07-27T01:21:58.727000", "content": "Hey, it's great to so see you guys grouping up
\n\nBest of luck on this competition!
", "votes": 1, "replies": [] }, { "id": 598986, "author_name": "Stanislav Blinov", "author_url": "", "post_date": "2019-08-14T10:28:51.090000", "content": "Hi guys! @muhakabartay @yassinealouini \nI'm kinda back into business (a little later than expected tho). Is your invitation still up?
", "votes": 1, "replies": [] }, { "id": 599064, "author_name": "Mukharbek Organokov", "author_url": "", "post_date": "2019-08-14T12:56:26.410000", "content": "Hi Stanislav @stanislavblinov
\n\nSome things changed but not vastly. I just a few days before I discussed with our team @yassinealouini and @harshit92 that we need someone with the model as you have. Because it can boost our score if you really reached -1.7 only via the usage of LGBM model. @yassinealouini @harshit92 what do you think of that if Stanislav will join us and two great and different models can be brought together?
\n\nP.S. @yassinealouini I told you (don't want to tell it obvious here how) that such exact -1.7 in LGBM can boost us. And of course, can boost Stanislav. but both teams able to fight for bronze I think (our silver is under big questions and with Stanislav's model we can fight for silver with higher confidence than now).
", "votes": 3, "replies": [] }, { "id": 599083, "author_name": "Stanislav Blinov", "author_url": "", "post_date": "2019-08-14T13:24:09.480000", "content": "@muhakabartay \nMy LGB-and-nothing-else model has CV -1.78 and LB -1.73.
\n\n\n\n\nwe can fight for silver
\n
Nah, green zone only :)
", "votes": 1, "replies": [] }, { "id": 599277, "author_name": "Mukharbek Organokov", "author_url": "", "post_date": "2019-08-14T18:42:21.767000", "content": "@stanislavblinov even better ))))))))))
\nLet's wait response of the other my teammates.
I have sent you a merge request. Hopefully our collaboration is fruitful.
", "votes": 2, "replies": [] }, { "id": 599639, "author_name": "Stanislav Blinov", "author_url": "", "post_date": "2019-08-15T07:59:19.027000", "content": "Accepted, thanks. \nAlso, I sent you an email
", "votes": 2, "replies": [] } ] }, { "id": 581370, "author_name": "Stanislav Blinov", "author_url": "", "post_date": "2019-07-21T20:45:10.980000", "content": "Hi, I'm interested!\nAbout me:\nI am quite new to kaggle though already earned a couple of medals, and not very experienced in data science in general (did my very first print('hello, world') ~7 months ago) and therefore would really love to have an experienced teammate / s.\nAbout my model:\nMy model is GPU-accelerated LGBM, mean CV -1.64, LB -1.57.\nImportant: I'm on vacation at the moment without access to my computer, therefore I won't be able to participate till 31st of July. I am ready to put some other things aside and make more spare time to spend on this comp in August (discussable 😉 )
\n\nAnyhow, if you're interested, you can contact me here on kaggle or via e-mail blinberg@mail.ru.
", "votes": 1, "replies": [ { "id": 581696, "author_name": "wayfarer", "author_url": "", "post_date": "2019-07-22T09:14:26.877000", "content": "I wish you have a great vacation and I hope you spend it happily!
", "votes": 4, "replies": [] } ] }, { "id": 581118, "author_name": "conradr71", "author_url": "", "post_date": "2019-07-21T12:54:32.120000", "content": "I can probably add some diversity to your models. I'm exclusively using graph cnns. My best LB score is from a single model trained on all coupling types.
", "votes": 1, "replies": [ { "id": 581140, "author_name": "Garrus", "author_url": "", "post_date": "2019-07-21T13:56:25.470000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 584938, "author_name": "", "author_url": "", "post_date": "2019-07-26T17:40:30.480000", "content": "", "votes": 0, "replies": [ { "id": 585111, "author_name": "Cozy Doomer", "author_url": "", "post_date": "2019-07-27T01:25:54.873000", "content": "We are using lightgbm and 2 different graph nn, the high score is from a combination of those \n(no single model above -1.9)
", "votes": 1, "replies": [] }, { "id": 588482, "author_name": "", "author_url": "", "post_date": "2019-07-30T16:19:40.167000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 581106, "author_name": "", "author_url": "", "post_date": "2019-07-21T12:38:44.930000", "content": "", "votes": 2, "replies": [ { "id": 581675, "author_name": "fwj7", "author_url": "", "post_date": "2019-07-22T08:43:04.550000", "content": "here
\n\nP.S: I am a fan of Robin VanPersi.
", "votes": 1, "replies": [] } ] }, { "id": 580783, "author_name": "", "author_url": "", "post_date": "2019-07-20T18:47:00.840000", "content": "", "votes": 0, "replies": [ { "id": 581035, "author_name": "Garrus", "author_url": "", "post_date": "2019-07-21T09:38:17.490000", "content": "We are looking for teammates with a better score than highest public kernel - 0.15
", "votes": 0, "replies": [] }, { "id": 584377, "author_name": "", "author_url": "", "post_date": "2019-07-25T20:56:55.730000", "content": "", "votes": 1, "replies": [] }, { "id": 588481, "author_name": "Mukharbek Organokov", "author_url": "", "post_date": "2019-07-30T16:18:56.653000", "content": "@doctorgg might join our team?
\n\n@yassinealouini someone from top80 might bring good advancements. What do you think?
", "votes": 1, "replies": [] }, { "id": 588535, "author_name": "Yassine Alouini", "author_url": "", "post_date": "2019-07-30T18:14:28.397000", "content": "Yes for sure, let me know if it is fine with you @doctorgg and I will send an invitation. :)
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "580504": "", "584944": "We try to form a team now with @yassinealouini \n@htopper @stanislavblinov would be might great to slightly increase the size of our team to be 4-5 in total (I pretty sure that more than 4-5 people in a team lead to degradation of efficiency). \nLet's see the opinion of @yassinealouini \nI stand with the more less same idea as @stanislavblinov, but to bring another method would be great since I don't think so that with LGBM or whatsoever we can even touch the LB Top100 ))). But neural networks power the top10 LB as we know.", "581370": "Hi, I'm interested!\n**About me:**\nI am quite new to kaggle though already earned a couple of medals, and not very experienced in data science in general (did my very first print('hello, world') ~7 months ago) and therefore would really love to have an experienced teammate / s.\n**About my model:**\nMy model is GPU-accelerated LGBM, mean CV -1.64, LB -1.57.\n**Important:** I'm on vacation at the moment without access to my computer, therefore I won't be able to participate till 31st of July. I am ready to put some other things aside and make more spare time to spend on this comp in August (discussable 😉 )\n\nAnyhow, if you're interested, you can contact me here on kaggle or via e-mail blinberg@mail.ru.", "581118": "I can probably add some diversity to your models. I'm exclusively using graph cnns. My best LB score is from a single model trained on all coupling types.", "584938": "", "581106": "", "580783": "" } }