I have a doubt, I noticed that the date in test set today is 2021-04-27, and training data is till 2021-04-30. Isn't that wrong, test date should be beyond the training set or I am missing something?
", "messages": [ { "id": "1349771", "postDate": "06/15/2021 04:18:20", "content": "I have a doubt, I noticed that the date in test set today is 2021-04-27, and training data is till 2021-04-30. Isn't that wrong, test date should be beyond the training set or I am missing something?
", "rawMarkdown": "I have a doubt, I noticed that the date in test set today is 2021-04-27, and training data is till 2021-04-30. Isn't that wrong, test date should be beyond the training set or I am missing something?", "votes": null }, { "id": "1350452", "postDate": "06/15/2021 12:58:10", "content": "The example_test.csv is just an example file that allows you to experiment with the mlb module locally. Right now, the test set used for the leaderboard is May 2021:
\n\nWhen you submit your notebook, it will be rerun on an unseen test set:
\nDuring the Training phase of the competition, this unseen test set is comprised of data for the month of May 2021 and the set of active players this year.
\n
See the data page for full details.
", "rawMarkdown": "The `example_test.csv` is just an example file that allows you to experiment with the mlb `module` locally. Right now, the test set used for the leaderboard is May 2021:\n\n> When you submit your notebook, it will be rerun on an unseen test set:\n> \n> During the Training phase of the competition, this unseen test set is comprised of data for the month of May 2021 and the set of active players this year.\n\nSee the data page for full details.", "votes": null }, { "id": "1350687", "postDate": "06/15/2021 16:51:29", "content": "When you say \"experiment with the mlb module locally\" I would love to assume that locally means I can experiment with my code on my local Linux machine - but I don't see that the \"mlb\" api is available to download.
\nIf anyone has been successful in a \"local\" bit of code to test thier submission process on the example files I would love to see it! I think I had some success in a previous competition that used a similiar process for feeding time data, but don't recall if that's a real memory or a wish nor can I recall the competition.
", "rawMarkdown": "When you say \"experiment with the mlb module locally\" I would love to assume that locally means I can experiment with my code on my local Linux machine - but I don't see that the \"mlb\" api is available to download.\n\nIf anyone has been successful in a \"local\" bit of code to test thier submission process on the example files I would love to see it! I think I had some success in a previous competition that used a similiar process for feeding time data, but don't recall if that's a real memory or a wish nor can I recall the competition.", "votes": null }, { "id": "1350699", "postDate": "06/15/2021 17:01:29", "content": "Riiid Answer Correctness Prediction
\nThis was the competition that used a similiar API and provided some example files - now - can I just find if my local use was a dream or reality!!!
", "rawMarkdown": "[Riiid Answer Correctness Prediction](https://www.kaggle.com/c/riiid-test-answer-prediction)\n\nThis was the competition that used a similiar API and provided some example files - now - can I just find if my local use was a dream or reality!!!", "votes": null }, { "id": "1350711", "postDate": "06/15/2021 17:11:16", "content": "OK - getting closer - when we download the entire data set to local machine an mlb folder is created with an .so file - but I am still not remembering how to install that file and use it (a benefit of being 74 years old is that every couple of years I can re-watch all the Seinfeld episodes and they seem new - clearly this is not a benefit in coding :)
", "rawMarkdown": "OK - getting closer - when we download the entire data set to local machine an mlb folder is created with an .so file - but I am still not remembering how to install that file and use it (a benefit of being 74 years old is that every couple of years I can re-watch all the Seinfeld episodes and they seem new - clearly this is not a benefit in coding :)", "votes": null }, { "id": "1350750", "postDate": "06/15/2021 17:59:11", "content": "To use locally you need to be running an identical version of Python as Kaggle notebooks (the module is compiled, hence requires exact version match) and then append the folder containing the mlb directory to your system path with sys.path.append(). If you're familiar with Docker you can match the Kaggle environment with our docker image (https://github.com/Kaggle/docker-python).
Thanks - I don't currently match Python versions and not real friends with Docker. I had located the data files on my local machine but my scripts are on my home server - so the attempt to import mlb failed but with the wrong reason:)
\nBut if I did match than on my local than I assume there would be nothing different I would need to do to run the mlb env on local vs within a kaggle kernel.
", "rawMarkdown": "Thanks - I don't currently match Python versions and not real friends with Docker. I had located the data files on my local machine but my scripts are on my home server - so the attempt to import mlb failed but with the wrong reason:)\n\nBut if I did match than on my local than I assume there would be nothing different I would need to do to run the mlb env on local vs within a kaggle kernel.", "votes": null }, { "id": "1359961", "postDate": "06/21/2021 17:27:34", "content": "Will - thanks for the tip that Python versions need to match. Did not do the Docker thing - just added a second environment to my Anaconda with Python 3.7 and the mlb now works like a charm.
", "rawMarkdown": "[Will](https://www.kaggle.com/wcukierski) - thanks for the tip that Python versions need to match. Did not do the Docker thing - just added a second environment to my Anaconda with Python 3.7 and the mlb now works like a charm.", "votes": null }, { "id": "1361370", "postDate": "06/22/2021 18:55:51", "content": "Can you share more details on how did you achieve it?
", "rawMarkdown": "Can you share more details on how did you achieve it?", "votes": null }, { "id": "1361546", "postDate": "06/22/2021 22:58:26", "content": "\nI am running local PC's with Ubuntu - using Anaconda installation - I reset the machines for this competition so they all had fresh OS and Anaconda. Which meant Python was version greater than 3.7.
\nThe mlb folder should be in the working directory - IF you have Python 3.7 it will function locally just like it does on Kaggle. You code needs to be like many of the shared kernels.
\nIf your asking how I got MLB to work - here's the short list.
\nconda install python=3.7
\nIn Aanaconda Navigator - click on environments - than Create+.
\nGive it a name - I did python37. You should get a drop down of version - select 3.7
\nYou need to activate the environment with each reboot.
\nconda activate python37
\nYou will need to install any libraries that you have been using in the new environment.
MLB will than feed the example_test data to your kernel - just like kaggle you can only run it once but a simple restart and clear does the trick. If you can't get it working than send me an email and we can work it out.
\nIf your asking how I got a successful submission generated I am not there yet - when I said \"works like a charm\" I simply meant that I can now get the same errors on my local machine as I was getting on Kaggle trying to create a submission that works.
", "rawMarkdown": "[EnricRovira](https://www.kaggle.com/enric1296)\n\nI am running local PC's with Ubuntu - using Anaconda installation - I reset the machines for this competition so they all had fresh OS and Anaconda. Which meant Python was version greater than 3.7. \n\nThe mlb folder should be in the working directory - IF you have Python 3.7 it will function locally just like it does on Kaggle. You code needs to be like many of the shared kernels.\n\nIf your asking how I got MLB to work - here's the short list.\n\nconda install python=3.7\nIn Aanaconda Navigator - click on environments - than Create+.\nGive it a name - I did python37. You should get a drop down of version - select 3.7\nYou need to activate the environment with each reboot.\nconda activate python37\nYou will need to install any libraries that you have been using in the new environment.\n\nMLB will than feed the example_test data to your kernel - just like kaggle you can only run it once but a simple restart and clear does the trick. If you can't get it working than send me an email and we can work it out.\n\n\n**If your asking how I got a successful submission** generated I am not there yet - when I said \"works like a charm\" I simply meant that I can now get the same errors on my local machine as I was getting on Kaggle trying to create a submission that works.", "votes": null }, { "id": "1375487", "postDate": "07/04/2021 08:49:21", "content": "Will June, July or August be used later?
", "rawMarkdown": "Will June, July or August be used later?", "votes": null }, { "id": "1375684", "postDate": "07/04/2021 12:17:06", "content": "Per other posts from the hosts - there will be optional new training set that includes June - we will have the option of using this additional. August will probably be part of the final private test set.
", "rawMarkdown": "Per other posts from the hosts - there will be optional new training set that includes June - we will have the option of using this additional. August will probably be part of the final private test set.", "votes": null } ], "comments": [ { "id": 1350452, "author_name": "wcukierski", "author_url": "", "post_date": "06/15/2021 12:58:10", "content": "The example_test.csv is just an example file that allows you to experiment with the mlb module locally. Right now, the test set used for the leaderboard is May 2021:
\n\nWhen you submit your notebook, it will be rerun on an unseen test set:
\nDuring the Training phase of the competition, this unseen test set is comprised of data for the month of May 2021 and the set of active players this year.
\n
See the data page for full details.
", "votes": null, "replies": [ { "id": 1350687, "author_name": "pcjimmmy", "author_url": "", "post_date": "06/15/2021 16:51:29", "content": "When you say \"experiment with the mlb module locally\" I would love to assume that locally means I can experiment with my code on my local Linux machine - but I don't see that the \"mlb\" api is available to download.
\nIf anyone has been successful in a \"local\" bit of code to test thier submission process on the example files I would love to see it! I think I had some success in a previous competition that used a similiar process for feeding time data, but don't recall if that's a real memory or a wish nor can I recall the competition.
", "votes": null, "replies": [] }, { "id": 1350699, "author_name": "pcjimmmy", "author_url": "", "post_date": "06/15/2021 17:01:29", "content": "Riiid Answer Correctness Prediction
\nThis was the competition that used a similiar API and provided some example files - now - can I just find if my local use was a dream or reality!!!
", "votes": null, "replies": [] }, { "id": 1350711, "author_name": "pcjimmmy", "author_url": "", "post_date": "06/15/2021 17:11:16", "content": "OK - getting closer - when we download the entire data set to local machine an mlb folder is created with an .so file - but I am still not remembering how to install that file and use it (a benefit of being 74 years old is that every couple of years I can re-watch all the Seinfeld episodes and they seem new - clearly this is not a benefit in coding :)
", "votes": null, "replies": [] }, { "id": 1350750, "author_name": "wcukierski", "author_url": "", "post_date": "06/15/2021 17:59:11", "content": "To use locally you need to be running an identical version of Python as Kaggle notebooks (the module is compiled, hence requires exact version match) and then append the folder containing the mlb directory to your system path with sys.path.append(). If you're familiar with Docker you can match the Kaggle environment with our docker image (https://github.com/Kaggle/docker-python).
Thanks - I don't currently match Python versions and not real friends with Docker. I had located the data files on my local machine but my scripts are on my home server - so the attempt to import mlb failed but with the wrong reason:)
\nBut if I did match than on my local than I assume there would be nothing different I would need to do to run the mlb env on local vs within a kaggle kernel.
", "votes": null, "replies": [] }, { "id": 1359961, "author_name": "pcjimmmy", "author_url": "", "post_date": "06/21/2021 17:27:34", "content": "Will - thanks for the tip that Python versions need to match. Did not do the Docker thing - just added a second environment to my Anaconda with Python 3.7 and the mlb now works like a charm.
", "votes": null, "replies": [] }, { "id": 1361370, "author_name": "enric1296", "author_url": "", "post_date": "06/22/2021 18:55:51", "content": "Can you share more details on how did you achieve it?
", "votes": null, "replies": [] }, { "id": 1361546, "author_name": "pcjimmmy", "author_url": "", "post_date": "06/22/2021 22:58:26", "content": "\nI am running local PC's with Ubuntu - using Anaconda installation - I reset the machines for this competition so they all had fresh OS and Anaconda. Which meant Python was version greater than 3.7.
\nThe mlb folder should be in the working directory - IF you have Python 3.7 it will function locally just like it does on Kaggle. You code needs to be like many of the shared kernels.
\nIf your asking how I got MLB to work - here's the short list.
\nconda install python=3.7
\nIn Aanaconda Navigator - click on environments - than Create+.
\nGive it a name - I did python37. You should get a drop down of version - select 3.7
\nYou need to activate the environment with each reboot.
\nconda activate python37
\nYou will need to install any libraries that you have been using in the new environment.
MLB will than feed the example_test data to your kernel - just like kaggle you can only run it once but a simple restart and clear does the trick. If you can't get it working than send me an email and we can work it out.
\nIf your asking how I got a successful submission generated I am not there yet - when I said \"works like a charm\" I simply meant that I can now get the same errors on my local machine as I was getting on Kaggle trying to create a submission that works.
", "votes": null, "replies": [] } ] }, { "id": 1375487, "author_name": "neverstoppredicting", "author_url": "", "post_date": "07/04/2021 08:49:21", "content": "Will June, July or August be used later?
", "votes": null, "replies": [ { "id": 1375684, "author_name": "pcjimmmy", "author_url": "", "post_date": "07/04/2021 12:17:06", "content": "Per other posts from the hosts - there will be optional new training set that includes June - we will have the option of using this additional. August will probably be part of the final private test set.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1349771": "I have a doubt, I noticed that the date in test set today is 2021-04-27, and training data is till 2021-04-30. Isn't that wrong, test date should be beyond the training set or I am missing something?", "1350452": "The `example_test.csv` is just an example file that allows you to experiment with the mlb `module` locally. Right now, the test set used for the leaderboard is May 2021:\n\n> When you submit your notebook, it will be rerun on an unseen test set:\n> \n> During the Training phase of the competition, this unseen test set is comprised of data for the month of May 2021 and the set of active players this year.\n\nSee the data page for full details.", "1350687": "When you say \"experiment with the mlb module locally\" I would love to assume that locally means I can experiment with my code on my local Linux machine - but I don't see that the \"mlb\" api is available to download.\n\nIf anyone has been successful in a \"local\" bit of code to test thier submission process on the example files I would love to see it! I think I had some success in a previous competition that used a similiar process for feeding time data, but don't recall if that's a real memory or a wish nor can I recall the competition.", "1350699": "[Riiid Answer Correctness Prediction](https://www.kaggle.com/c/riiid-test-answer-prediction)\n\nThis was the competition that used a similiar API and provided some example files - now - can I just find if my local use was a dream or reality!!!", "1350711": "OK - getting closer - when we download the entire data set to local machine an mlb folder is created with an .so file - but I am still not remembering how to install that file and use it (a benefit of being 74 years old is that every couple of years I can re-watch all the Seinfeld episodes and they seem new - clearly this is not a benefit in coding :)", "1350750": "To use locally you need to be running an identical version of Python as Kaggle notebooks (the module is compiled, hence requires exact version match) and then append the folder containing the mlb directory to your system path with `sys.path.append()`. If you're familiar with Docker you can match the Kaggle environment with our docker image (https://github.com/Kaggle/docker-python).", "1350769": "Thanks - I don't currently match Python versions and not real friends with Docker. I had located the data files on my local machine but my scripts are on my home server - so the attempt to import mlb failed but with the wrong reason:)\n\nBut if I did match than on my local than I assume there would be nothing different I would need to do to run the mlb env on local vs within a kaggle kernel.", "1359961": "[Will](https://www.kaggle.com/wcukierski) - thanks for the tip that Python versions need to match. Did not do the Docker thing - just added a second environment to my Anaconda with Python 3.7 and the mlb now works like a charm.", "1361370": "Can you share more details on how did you achieve it?", "1361546": "[EnricRovira](https://www.kaggle.com/enric1296)\n\nI am running local PC's with Ubuntu - using Anaconda installation - I reset the machines for this competition so they all had fresh OS and Anaconda. Which meant Python was version greater than 3.7. \n\nThe mlb folder should be in the working directory - IF you have Python 3.7 it will function locally just like it does on Kaggle. You code needs to be like many of the shared kernels.\n\nIf your asking how I got MLB to work - here's the short list.\n\nconda install python=3.7\nIn Aanaconda Navigator - click on environments - than Create+.\nGive it a name - I did python37. You should get a drop down of version - select 3.7\nYou need to activate the environment with each reboot.\nconda activate python37\nYou will need to install any libraries that you have been using in the new environment.\n\nMLB will than feed the example_test data to your kernel - just like kaggle you can only run it once but a simple restart and clear does the trick. If you can't get it working than send me an email and we can work it out.\n\n\n**If your asking how I got a successful submission** generated I am not there yet - when I said \"works like a charm\" I simply meant that I can now get the same errors on my local machine as I was getting on Kaggle trying to create a submission that works.", "1375487": "Will June, July or August be used later?", "1375684": "Per other posts from the hosts - there will be optional new training set that includes June - we will have the option of using this additional. August will probably be part of the final private test set." }, "source": "meta" }