Why is there no Kernel tab for the competition?
\n\nI thought the goal was to spread knowledge and awereness of the workshop, the dataset and new models for them - allowing users to post their code in Kernels would help a lot.
", "messages": [ { "id": 188086, "postDate": "2017-06-02T00:00:23.967Z", "content": "Why is there no Kernel tab for the competition?
\n\nI thought the goal was to spread knowledge and awereness of the workshop, the dataset and new models for them - allowing users to post their code in Kernels would help a lot.
", "rawMarkdown": "Why is there no Kernel tab for the competition?\n\n\nI thought the goal was to spread knowledge and awereness of the workshop, the dataset and new models for them - allowing users to post their code in Kernels would help a lot.", "votes": 12 }, { "id": 188381, "postDate": "2017-06-02T18:58:04.610Z", "content": "@FrankHerfert,
\n\nWe decided since the data is too large for Kaggle Kernels, we wouldn't enable Kernels in this competition. If you think it's still useful with a subset of the images to run on Kernels, let us know and I can get it enabled.
", "rawMarkdown": "@FrankHerfert,\n\nWe decided since the data is too large for Kaggle Kernels, we wouldn't enable Kernels in this competition. If you think it's still useful with a subset of the images to run on Kernels, let us know and I can get it enabled.", "votes": 2, "replies": [ { "id": 188451, "postDate": "2017-06-02T22:14:44.373Z", "content": "Thanks for the info.
\n\nThe dataset size is a good argument and having people constantly run into Kernel limitations might also not be very encouraging.
\n\nOn the other hand, a few other recent competitions had datasets that were also quite large and/or required computational power that far exceeded the Kaggle Kernel limits still offered users to share their code via Kernels. For the Quora duplicate questions challenge, you can't really process the whole dataset in the Kernel, too, but it's a great way to share ideas and insights. Especially if you run a notebook locally, then upload it to Kaggle, you can easily share code and results without using Kaggle's computing resources.
\n\nFor me, Kernels are more about spreading and finding ideas than having a cloud-based platform to run my code on. It's more comfortable than sharing a link to github, too, and enables discussion about the code right on Kaggle.
\n\nMaybe enable Kernels with a small subset of images or just as a way to upload notebooks/scripts without any computational power behind it. This might make it easier for people to start working on the dataset.
", "rawMarkdown": "Thanks for the info.\n\nThe dataset size is a good argument and having people constantly run into Kernel limitations might also not be very encouraging.\n\nOn the other hand, a few other recent competitions had datasets that were also quite large and/or required computational power that far exceeded the Kaggle Kernel limits still offered users to share their code via Kernels. For the Quora duplicate questions challenge, you can't really process the whole dataset in the Kernel, too, but it's a great way to share ideas and insights. Especially if you run a notebook locally, then upload it to Kaggle, you can easily share code and results without using Kaggle's computing resources. \n\nFor me, Kernels are more about spreading and finding ideas than having a cloud-based platform to run my code on. It's more comfortable than sharing a link to github, too, and enables discussion about the code right on Kaggle.\n\nMaybe enable Kernels with a small subset of images or just as a way to upload notebooks/scripts without any computational power behind it. This might make it easier for people to start working on the dataset.\n\n", "votes": 10 }, { "id": 189407, "postDate": "2017-06-06T00:32:13.163Z", "content": "Kernels enabled with sample images.
", "rawMarkdown": "Kernels enabled with sample images. ", "votes": 1 }, { "id": 191505, "postDate": "2017-06-10T16:30:55.183Z", "content": "Hi Wendy,
\nThanks for turning on kernels. However, when I try creating a new kernel, I am met with the \"Oops, something went wrong\" page: https://www.kaggle.com/anokas/kernels/scripts/new?competitionId=6538
\nAm I doing something wrong?
We are checking into the Kernels issue, thanks for your feedback!
", "rawMarkdown": "We are checking into the Kernels issue, thanks for your feedback! " } ] } ], "comments": [ { "id": 188381, "author_name": "Wendy Kan", "author_url": "", "post_date": "2017-06-02T18:58:04.610000", "content": "@FrankHerfert,
\n\nWe decided since the data is too large for Kaggle Kernels, we wouldn't enable Kernels in this competition. If you think it's still useful with a subset of the images to run on Kernels, let us know and I can get it enabled.
", "votes": 2, "replies": [ { "id": 188451, "author_name": "FrankHerfert", "author_url": "", "post_date": "2017-06-02T22:14:44.373000", "content": "Thanks for the info.
\n\nThe dataset size is a good argument and having people constantly run into Kernel limitations might also not be very encouraging.
\n\nOn the other hand, a few other recent competitions had datasets that were also quite large and/or required computational power that far exceeded the Kaggle Kernel limits still offered users to share their code via Kernels. For the Quora duplicate questions challenge, you can't really process the whole dataset in the Kernel, too, but it's a great way to share ideas and insights. Especially if you run a notebook locally, then upload it to Kaggle, you can easily share code and results without using Kaggle's computing resources.
\n\nFor me, Kernels are more about spreading and finding ideas than having a cloud-based platform to run my code on. It's more comfortable than sharing a link to github, too, and enables discussion about the code right on Kaggle.
\n\nMaybe enable Kernels with a small subset of images or just as a way to upload notebooks/scripts without any computational power behind it. This might make it easier for people to start working on the dataset.
", "votes": 10, "replies": [] }, { "id": 189407, "author_name": "Wendy Kan", "author_url": "", "post_date": "2017-06-06T00:32:13.163000", "content": "Kernels enabled with sample images.
", "votes": 1, "replies": [] }, { "id": 191505, "author_name": "anokas", "author_url": "", "post_date": "2017-06-10T16:30:55.183000", "content": "Hi Wendy,
\nThanks for turning on kernels. However, when I try creating a new kernel, I am met with the \"Oops, something went wrong\" page: https://www.kaggle.com/anokas/kernels/scripts/new?competitionId=6538
\nAm I doing something wrong?
We are checking into the Kernels issue, thanks for your feedback!
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "188086": "Why is there no Kernel tab for the competition?\n\n\nI thought the goal was to spread knowledge and awereness of the workshop, the dataset and new models for them - allowing users to post their code in Kernels would help a lot.", "188381": "@FrankHerfert,\n\nWe decided since the data is too large for Kaggle Kernels, we wouldn't enable Kernels in this competition. If you think it's still useful with a subset of the images to run on Kernels, let us know and I can get it enabled." } }