New to machine learning and data science? No question is too basic or too simple. Feel free to start your own thread, or use this thread as a place to post any first-timer clarifying questions for the Kaggle community to help you with!
\n\nIf you would consider yourself a beginner but don't know where to get started, let other Kagglers help you take your first steps here!
\n\nNew to Kaggle? Take a look at a few videos our very own Dr. Rachael Tatman has put together to learn a bit more about site etiquette, Kaggle lingo, and how to enter a competition using Kaggle Kernels (Notebooks).
", "messages": [ { "id": "625255", "postDate": "09/12/2019 20:37:20", "content": "New to machine learning and data science? No question is too basic or too simple. Feel free to start your own thread, or use this thread as a place to post any first-timer clarifying questions for the Kaggle community to help you with!
\n\nIf you would consider yourself a beginner but don't know where to get started, let other Kagglers help you take your first steps here!
\n\nNew to Kaggle? Take a look at a few videos our very own Dr. Rachael Tatman has put together to learn a bit more about site etiquette, Kaggle lingo, and how to enter a competition using Kaggle Kernels (Notebooks).
", "rawMarkdown": "New to machine learning and data science? No question is too basic or too simple. Feel free to start your own thread, or use this thread as a place to post any first-timer clarifying questions for the Kaggle community to help you with! \n\nIf you would consider yourself a beginner but don't know where to get started, let other Kagglers help you take your first steps here!\n\nNew to Kaggle? Take a look at a few videos our very own Dr. Rachael Tatman has put together to learn a bit more about [site etiquette](https://www.youtube.com/watch?v=aIus8si_Et0), [Kaggle lingo](https://www.youtube.com/watch?v=sEJHyuWKd-s), and [how to enter a competition using Kaggle Kernels (Notebooks)](https://www.youtube.com/watch?&v=GJBOMWpLpTQ).", "votes": null }, { "id": "625809", "postDate": "09/13/2019 13:32:18", "content": "Hello everyone! I am new to Kaggle. In fact, this would be my first competition. I have a very basic doubt which is can we work (train + test) on the Kaggle kernels itself because the data is quite large and I do not have a strong enough local machine.
", "rawMarkdown": "Hello everyone! I am new to Kaggle. In fact, this would be my first competition. I have a very basic doubt which is can we work (train + test) on the Kaggle kernels itself because the data is quite large and I do not have a strong enough local machine.", "votes": null }, { "id": "626204", "postDate": "09/14/2019 02:40:18", "content": "Thank you for your encouragement! Also new.
", "rawMarkdown": "Thank you for your encouragement! Also new.", "votes": null }, { "id": "627445", "postDate": "09/16/2019 01:50:29", "content": "Thanks Me to New..
", "rawMarkdown": "Thanks Me to New..", "votes": null }, { "id": "627519", "postDate": "09/16/2019 05:05:31", "content": "Hi Rohil. I haven't started the competition yet, so I am not sure if using all the available data for training would exceed the kernel time limit, but if that happens, you can train your model using GCP free tier which includes a 12-month free trial with $300 credit to use with any GCP services. P.S. I think you should use Kaggle kernel first, and only use a small subset of the data to debug your pipeline and to experiments with different tricks and models. Hope this help and best of luck!
", "rawMarkdown": "Hi Rohil. I haven't started the competition yet, so I am not sure if using all the available data for training would exceed the kernel time limit, but if that happens, you can train your model using GCP free tier which includes a 12-month free trial with $300 credit to use with any GCP services. P.S. I think you should use Kaggle kernel first, and only use a small subset of the data to debug your pipeline and to experiments with different tricks and models. Hope this help and best of luck!", "votes": null }, { "id": "632873", "postDate": "09/24/2019 07:05:40", "content": "Thanks, Michael for the great advice.
", "rawMarkdown": "Thanks, Michael for the great advice.", "votes": null }, { "id": "634180", "postDate": "09/26/2019 01:56:50", "content": "Hi! This is my first Kaggle competition. I'm in my last year of a Computer Science MS, and I have to pick a project for my parallel processing course. This seems like an ambitious project, so I'm looking forward to learning a lot! I have worked a little with a few different types of CNNs, like U-Net and ResNet, but no experience working with 3D or temporal modelling, so I'm looking forward to reading about other people's approaches to this problem. If anyone has a good suggestion on where to start research-wise, I would be very grateful for any suggestions!\n- Dan McGonigle
", "rawMarkdown": "Hi! This is my first Kaggle competition. I'm in my last year of a Computer Science MS, and I have to pick a project for my parallel processing course. This seems like an ambitious project, so I'm looking forward to learning a lot! I have worked a little with a few different types of CNNs, like U-Net and ResNet, but no experience working with 3D or temporal modelling, so I'm looking forward to reading about other people's approaches to this problem. If anyone has a good suggestion on where to start research-wise, I would be very grateful for any suggestions!\n- Dan McGonigle", "votes": null }, { "id": "654497", "postDate": "10/21/2019 23:51:08", "content": "Hey Everyone!\nThis is my first Kaggle compettion.\nI am joining in very late and I have a lot of ground to cover up.\nBut I hope to learn a lot in the process.\nI am open to any suggestions people have that can help me get on track.
\n\nThank you,\nRishabh
", "rawMarkdown": "Hey Everyone!\nThis is my first Kaggle compettion.\nI am joining in very late and I have a lot of ground to cover up.\nBut I hope to learn a lot in the process.\nI am open to any suggestions people have that can help me get on track.\n\nThank you,\nRishabh", "votes": null }, { "id": "655178", "postDate": "10/22/2019 19:03:57", "content": "Good luck! and you are already doing great!
", "rawMarkdown": "Good luck! and you are already doing great!", "votes": null }, { "id": "659634", "postDate": "10/28/2019 02:22:59", "content": "Being an absolute noob in machine learning and kaggle. I wanted to take the biggest, baddest competition in the yard and knock it down. Hence Lyft3d.
\n\nBut just like Rohil, I am stuck without enough local storage. These appear to be the options;
\n\nGCP / AWS free tier - already exhausted on previous work - not an option for me.
Use external hard disk - any issues with performance?
Write script to download X% of training and test data (read key file - sample X % - download lidar & jpegs for selected keys) - Are there any methods / scripts already out there?
Advice / experience sharing would be much appreciated.
", "rawMarkdown": "Being an absolute noob in machine learning and kaggle. I wanted to take the biggest, baddest competition in the yard and knock it down. Hence Lyft3d. \n\nBut just like Rohil, I am stuck without enough local storage. These appear to be the options;\n\n1. GCP / AWS free tier - already exhausted on previous work - not an option for me. \n\n2. Use external hard disk - any issues with performance? \n\n3. Write script to download X% of training and test data (read key file - sample X % - download lidar & jpegs for selected keys) - Are there any methods / scripts already out there?\n\nAdvice / experience sharing would be much appreciated.", "votes": null } ], "comments": [ { "id": 625809, "author_name": "thedatamonk", "author_url": "", "post_date": "09/13/2019 13:32:18", "content": "Hello everyone! I am new to Kaggle. In fact, this would be my first competition. I have a very basic doubt which is can we work (train + test) on the Kaggle kernels itself because the data is quite large and I do not have a strong enough local machine.
", "votes": null, "replies": [ { "id": 627519, "author_name": "lipeng2", "author_url": "", "post_date": "09/16/2019 05:05:31", "content": "Hi Rohil. I haven't started the competition yet, so I am not sure if using all the available data for training would exceed the kernel time limit, but if that happens, you can train your model using GCP free tier which includes a 12-month free trial with $300 credit to use with any GCP services. P.S. I think you should use Kaggle kernel first, and only use a small subset of the data to debug your pipeline and to experiments with different tricks and models. Hope this help and best of luck!
", "votes": null, "replies": [] }, { "id": 632873, "author_name": "thedatamonk", "author_url": "", "post_date": "09/24/2019 07:05:40", "content": "Thanks, Michael for the great advice.
", "votes": null, "replies": [] }, { "id": 659634, "author_name": "mpdroid", "author_url": "", "post_date": "10/28/2019 02:22:59", "content": "Being an absolute noob in machine learning and kaggle. I wanted to take the biggest, baddest competition in the yard and knock it down. Hence Lyft3d.
\n\nBut just like Rohil, I am stuck without enough local storage. These appear to be the options;
\n\nGCP / AWS free tier - already exhausted on previous work - not an option for me.
Use external hard disk - any issues with performance?
Write script to download X% of training and test data (read key file - sample X % - download lidar & jpegs for selected keys) - Are there any methods / scripts already out there?
Advice / experience sharing would be much appreciated.
", "votes": null, "replies": [] } ] }, { "id": 626204, "author_name": "rur4893", "author_url": "", "post_date": "09/14/2019 02:40:18", "content": "Thank you for your encouragement! Also new.
", "votes": null, "replies": [] }, { "id": 627445, "author_name": "sureshalex", "author_url": "", "post_date": "09/16/2019 01:50:29", "content": "Thanks Me to New..
", "votes": null, "replies": [] }, { "id": 634180, "author_name": "dpmcgonigle", "author_url": "", "post_date": "09/26/2019 01:56:50", "content": "Hi! This is my first Kaggle competition. I'm in my last year of a Computer Science MS, and I have to pick a project for my parallel processing course. This seems like an ambitious project, so I'm looking forward to learning a lot! I have worked a little with a few different types of CNNs, like U-Net and ResNet, but no experience working with 3D or temporal modelling, so I'm looking forward to reading about other people's approaches to this problem. If anyone has a good suggestion on where to start research-wise, I would be very grateful for any suggestions!\n- Dan McGonigle
", "votes": null, "replies": [] }, { "id": 654497, "author_name": "rishabhgks19", "author_url": "", "post_date": "10/21/2019 23:51:08", "content": "Hey Everyone!\nThis is my first Kaggle compettion.\nI am joining in very late and I have a lot of ground to cover up.\nBut I hope to learn a lot in the process.\nI am open to any suggestions people have that can help me get on track.
\n\nThank you,\nRishabh
", "votes": null, "replies": [ { "id": 655178, "author_name": "drnium", "author_url": "", "post_date": "10/22/2019 19:03:57", "content": "Good luck! and you are already doing great!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "625255": "New to machine learning and data science? No question is too basic or too simple. Feel free to start your own thread, or use this thread as a place to post any first-timer clarifying questions for the Kaggle community to help you with! \n\nIf you would consider yourself a beginner but don't know where to get started, let other Kagglers help you take your first steps here!\n\nNew to Kaggle? Take a look at a few videos our very own Dr. Rachael Tatman has put together to learn a bit more about [site etiquette](https://www.youtube.com/watch?v=aIus8si_Et0), [Kaggle lingo](https://www.youtube.com/watch?v=sEJHyuWKd-s), and [how to enter a competition using Kaggle Kernels (Notebooks)](https://www.youtube.com/watch?&v=GJBOMWpLpTQ).", "625809": "Hello everyone! I am new to Kaggle. In fact, this would be my first competition. I have a very basic doubt which is can we work (train + test) on the Kaggle kernels itself because the data is quite large and I do not have a strong enough local machine.", "626204": "Thank you for your encouragement! Also new.", "627445": "Thanks Me to New..", "627519": "Hi Rohil. I haven't started the competition yet, so I am not sure if using all the available data for training would exceed the kernel time limit, but if that happens, you can train your model using GCP free tier which includes a 12-month free trial with $300 credit to use with any GCP services. P.S. I think you should use Kaggle kernel first, and only use a small subset of the data to debug your pipeline and to experiments with different tricks and models. Hope this help and best of luck!", "632873": "Thanks, Michael for the great advice.", "634180": "Hi! This is my first Kaggle competition. I'm in my last year of a Computer Science MS, and I have to pick a project for my parallel processing course. This seems like an ambitious project, so I'm looking forward to learning a lot! I have worked a little with a few different types of CNNs, like U-Net and ResNet, but no experience working with 3D or temporal modelling, so I'm looking forward to reading about other people's approaches to this problem. If anyone has a good suggestion on where to start research-wise, I would be very grateful for any suggestions!\n- Dan McGonigle", "654497": "Hey Everyone!\nThis is my first Kaggle compettion.\nI am joining in very late and I have a lot of ground to cover up.\nBut I hope to learn a lot in the process.\nI am open to any suggestions people have that can help me get on track.\n\nThank you,\nRishabh", "655178": "Good luck! and you are already doing great!", "659634": "Being an absolute noob in machine learning and kaggle. I wanted to take the biggest, baddest competition in the yard and knock it down. Hence Lyft3d. \n\nBut just like Rohil, I am stuck without enough local storage. These appear to be the options;\n\n1. GCP / AWS free tier - already exhausted on previous work - not an option for me. \n\n2. Use external hard disk - any issues with performance? \n\n3. Write script to download X% of training and test data (read key file - sample X % - download lidar & jpegs for selected keys) - Are there any methods / scripts already out there?\n\nAdvice / experience sharing would be much appreciated." }, "source": "meta" }