Even though the problem looks straightforward, it's not easy to approach this dataset. I tried to compile a starter pack for that reason.
\nlibvips/pyvips Installation and Getting Started
\nThis notebook shows how to install and use libvips in an offline environment. It allows you to resize images that doesn't fit into memory. This was a requirement when GPU notebook's RAM was 13 GBs but all of the images can be read on GPU notebooks now since it is increased to 29 GBs.
UBC-OCEAN -JPEG Dataset Pipeline
\nThis notebook shows how to export images as JPEGs and thus download them in a shorter amount of time.
UBC-OCEAN - EDA
\nThis is an EDA notebook on which I explored basic properties of the competition dataset and I visualized all of the images so you don't have to do it one by one.
Next step is implementing a training/inference pipeline and submit predictions to Kaggle. Good luck with the competition!
", "messages": [ { "id": 2489862, "postDate": "2023-10-20T09:03:03.353Z", "content": "Even though the problem looks straightforward, it's not easy to approach this dataset. I tried to compile a starter pack for that reason.
\nlibvips/pyvips Installation and Getting Started
\nThis notebook shows how to install and use libvips in an offline environment. It allows you to resize images that doesn't fit into memory. This was a requirement when GPU notebook's RAM was 13 GBs but all of the images can be read on GPU notebooks now since it is increased to 29 GBs.
UBC-OCEAN -JPEG Dataset Pipeline
\nThis notebook shows how to export images as JPEGs and thus download them in a shorter amount of time.
UBC-OCEAN - EDA
\nThis is an EDA notebook on which I explored basic properties of the competition dataset and I visualized all of the images so you don't have to do it one by one.
Next step is implementing a training/inference pipeline and submit predictions to Kaggle. Good luck with the competition!
", "rawMarkdown": "Even though the problem looks straightforward, it's not easy to approach this dataset. I tried to compile a starter pack for that reason.\n\n[libvips/pyvips Installation and Getting Started](https://www.kaggle.com/code/gunesevitan/libvips-pyvips-installation-and-getting-started)\nThis notebook shows how to install and use libvips in an offline environment. It allows you to resize images that doesn't fit into memory. This was a requirement when GPU notebook's RAM was 13 GBs but all of the images can be read on GPU notebooks now since it is increased to 29 GBs.\n\n[UBC-OCEAN -JPEG Dataset Pipeline](https://www.kaggle.com/code/gunesevitan/ubc-ocean-jpeg-dataset-pipeline)\nThis notebook shows how to export images as JPEGs and thus download them in a shorter amount of time.\n\n[UBC-OCEAN - EDA](https://www.kaggle.com/code/gunesevitan/ubc-ocean-eda)\nThis is an EDA notebook on which I explored basic properties of the competition dataset and I visualized all of the images so you don't have to do it one by one.\n\nNext step is implementing a training/inference pipeline and submit predictions to Kaggle. Good luck with the competition!", "votes": 27 }, { "id": 2492595, "postDate": "2023-10-22T16:49:44.670Z", "content": "Really a great job.Thank you for all your works which helped me a lot as a new kaggler and a college freshman.
", "rawMarkdown": "Really a great job.Thank you for all your works which helped me a lot as a new kaggler and a college freshman.", "votes": 1 }, { "id": 2507778, "postDate": "2023-11-01T08:35:11.603Z", "content": "Thank you for providing this useful resource. Just joined the competition, will help me a great deal!
", "rawMarkdown": "Thank you for providing this useful resource. Just joined the competition, will help me a great deal!", "votes": 2 }, { "id": 2507532, "postDate": "2023-11-01T04:59:47.463Z", "content": "I think that after starter you can also check baseline with Lightning⚡TIMM scores 0.4+ on LB become TOP 5%
", "rawMarkdown": "I think that after starter you can also check [baseline with Lightning⚡TIMM scores 0.4+ on LB become TOP 5%](https://www.kaggle.com/competitions/UBC-OCEAN/discussion/452165)", "votes": 2 } ], "comments": [ { "id": 2492595, "author_name": "LuoZiqian", "author_url": "", "post_date": "2023-10-22T16:49:44.670000", "content": "Really a great job.Thank you for all your works which helped me a lot as a new kaggler and a college freshman.
", "votes": 1, "replies": [] }, { "id": 2507778, "author_name": "TensorKitty", "author_url": "", "post_date": "2023-11-01T08:35:11.603000", "content": "Thank you for providing this useful resource. Just joined the competition, will help me a great deal!
", "votes": 2, "replies": [] }, { "id": 2507532, "author_name": "Jirka", "author_url": "", "post_date": "2023-11-01T04:59:47.463000", "content": "I think that after starter you can also check baseline with Lightning⚡TIMM scores 0.4+ on LB become TOP 5%
", "votes": 2, "replies": [] } ], "raw_markdown_by_id": { "2489862": "Even though the problem looks straightforward, it's not easy to approach this dataset. I tried to compile a starter pack for that reason.\n\n[libvips/pyvips Installation and Getting Started](https://www.kaggle.com/code/gunesevitan/libvips-pyvips-installation-and-getting-started)\nThis notebook shows how to install and use libvips in an offline environment. It allows you to resize images that doesn't fit into memory. This was a requirement when GPU notebook's RAM was 13 GBs but all of the images can be read on GPU notebooks now since it is increased to 29 GBs.\n\n[UBC-OCEAN -JPEG Dataset Pipeline](https://www.kaggle.com/code/gunesevitan/ubc-ocean-jpeg-dataset-pipeline)\nThis notebook shows how to export images as JPEGs and thus download them in a shorter amount of time.\n\n[UBC-OCEAN - EDA](https://www.kaggle.com/code/gunesevitan/ubc-ocean-eda)\nThis is an EDA notebook on which I explored basic properties of the competition dataset and I visualized all of the images so you don't have to do it one by one.\n\nNext step is implementing a training/inference pipeline and submit predictions to Kaggle. Good luck with the competition!", "2492595": "Really a great job.Thank you for all your works which helped me a lot as a new kaggler and a college freshman.", "2507778": "Thank you for providing this useful resource. Just joined the competition, will help me a great deal!", "2507532": "I think that after starter you can also check [baseline with Lightning⚡TIMM scores 0.4+ on LB become TOP 5%](https://www.kaggle.com/competitions/UBC-OCEAN/discussion/452165)" } }