{ "id": 306152, "title": "5 Reduced Resolution Image Datasets (128, 256, 384, 512, 1024) 👔👖", "url": "/competitions/h-and-m-personalized-fashion-recommendations/discussion/306152", "author_name": "", "post_date": "2022-02-08T10:32:08.504292100Z", "votes": 100, "comment_count": 21, "views": 0, "content": "

Hello everyone!

\n

I have created 5 different datasets for this competition. I have taken the original images and scaled them down to 5 different lower resolutions (128 x 128, 256 x 256, 384 x 384, 512 x 512 and
\n1024 x1024 ). From previous competitions, we have seen that higher resolutions generally have better model performance but these provide a great starting point.

\n

All 5 datasets also include the 4 tabular CSV files in each of these datasets for convenience.

\n

These datasets are also a lot smaller in scale compared to the original dataset which should make it easier to download locally or even use in a Kaggle kernel.
\nPlease let me know if you have any feedback ✨

\n

Here are the dataset links:

\n\n

If you want are interested in knowing how the data was compressed and which interpolation method was used, check out this notebook - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256

", "messages": [ { "id": "1681218", "postDate": "02/08/2022 10:32:08", "content": "

Hello everyone!

\n

I have created 5 different datasets for this competition. I have taken the original images and scaled them down to 5 different lower resolutions (128 x 128, 256 x 256, 384 x 384, 512 x 512 and
\n1024 x1024 ). From previous competitions, we have seen that higher resolutions generally have better model performance but these provide a great starting point.

\n

All 5 datasets also include the 4 tabular CSV files in each of these datasets for convenience.

\n

These datasets are also a lot smaller in scale compared to the original dataset which should make it easier to download locally or even use in a Kaggle kernel.
\nPlease let me know if you have any feedback ✨

\n

Here are the dataset links:

\n\n

If you want are interested in knowing how the data was compressed and which interpolation method was used, check out this notebook - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256

", "rawMarkdown": "Hello everyone! \n\nI have created 5 different datasets for this competition. I have taken the original images and scaled them down to 5 different lower resolutions (128 x 128, 256 x 256, 384 x 384, 512 x 512 and \n1024 x1024 ). From previous competitions, we have seen that higher resolutions generally have better model performance but these provide a great starting point.\n\nAll 5 datasets also include the `4 tabular CSV` files in each of these datasets for convenience.\n\nThese datasets are also a lot smaller in scale compared to the original dataset which should make it easier to download locally or even use in a Kaggle kernel. \nPlease let me know if you have any feedback ✨\n\nHere are the dataset links:\n\n* (128 x 128 dataset) - https://www.kaggle.com/odins0n/handm-dataset-128x128\n* (256 x 256 dataset) - https://www.kaggle.com/odins0n/hm256x256\n* (384 x 384 dataset) - https://www.kaggle.com/odins0n/hm384x384\n* (512 x 512 dataset) - https://www.kaggle.com/odins0n/hm512x512\n* (1024 x 1024 dataset) - https://www.kaggle.com/odins0n/hm1024x1024\n\nIf you want are interested in knowing how the data was compressed and which interpolation method was used, check out this notebook - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256", "votes": null }, { "id": "1681672", "postDate": "02/08/2022 16:13:49", "content": "

Thanks for sharing @odins0n

", "rawMarkdown": "Thanks for sharing @odins0n", "votes": null }, { "id": "1681686", "postDate": "02/08/2022 16:26:58", "content": "

Welcome @subhamjain

", "rawMarkdown": "Welcome @subhamjain", "votes": null }, { "id": "1682322", "postDate": "02/09/2022 04:51:41", "content": "

Thanks a lot for sharing @odins0n
\nI was looking for 256x256. Any chance we can get 224x224 ?

", "rawMarkdown": "Thanks a lot for sharing @odins0n \nI was looking for 256x256. Any chance we can get 224x224 ?", "votes": null }, { "id": "1682424", "postDate": "02/09/2022 06:53:07", "content": "

The dataset size reduced significantly. Thanks for sharing

", "rawMarkdown": "The dataset size reduced significantly. Thanks for sharing", "votes": null }, { "id": "1682575", "postDate": "02/09/2022 08:52:24", "content": "

Welcome @rahulshastri1

", "rawMarkdown": "Welcome @rahulshastri1", "votes": null }, { "id": "1682576", "postDate": "02/09/2022 08:53:06", "content": "

Welcome @sanjeevprajapati2
\nYou can use this script to generate a dataset of any dimension you want - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256

", "rawMarkdown": "Welcome @sanjeevprajapati2 \nYou can use this script to generate a dataset of any dimension you want - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256", "votes": null }, { "id": "1695822", "postDate": "02/18/2022 11:41:15", "content": "

Thank you so much for this.

", "rawMarkdown": "Thank you so much for this.", "votes": null }, { "id": "1695969", "postDate": "02/18/2022 14:06:56", "content": "

Welcome @paweljankiewicz

", "rawMarkdown": "Welcome @paweljankiewicz", "votes": null }, { "id": "1698090", "postDate": "02/20/2022 06:11:10", "content": "

Thanks for creating these @odins0n!

\n

Going through the images, they look like perfectly centered product images.

\n

I suspect we can apply center cropping safely without any issues-maybe an idea for v2 of your datasets? :)

", "rawMarkdown": "Thanks for creating these @odins0n! \n\nGoing through the images, they look like perfectly centered product images. \n\nI suspect we can apply center cropping safely without any issues-maybe an idea for v2 of your datasets? :)", "votes": null }, { "id": "1698155", "postDate": "02/20/2022 07:24:18", "content": "

Welcome @init27
\nThat's a great idea. I will surely try out centre cropping and check for any difference in results.

", "rawMarkdown": "Welcome @init27 \nThat's a great idea. I will surely try out centre cropping and check for any difference in results.", "votes": null }, { "id": "1698673", "postDate": "02/20/2022 15:13:40", "content": "

Thank you for spending some time doing it for us. I appreciate it, @odins0n!

", "rawMarkdown": "Thank you for spending some time doing it for us. I appreciate it, @odins0n!", "votes": null }, { "id": "1698686", "postDate": "02/20/2022 15:23:09", "content": "

Welcome @vad13irt

", "rawMarkdown": "Welcome @vad13irt", "votes": null }, { "id": "1702148", "postDate": "02/23/2022 11:20:56", "content": "

Thank you very much for sharing @odins0n
\nReally helpful! Always good ideas 🧐

", "rawMarkdown": "Thank you very much for sharing @odins0n \nReally helpful! Always good ideas 🧐", "votes": null }, { "id": "1702188", "postDate": "02/23/2022 11:50:12", "content": "

Welcome @datascientistfp

", "rawMarkdown": "Welcome @datascientistfp", "votes": null }, { "id": "1705958", "postDate": "02/27/2022 02:45:26", "content": "

WOW Thanks for sharing !

", "rawMarkdown": "WOW Thanks for sharing !", "votes": null }, { "id": "1707015", "postDate": "02/28/2022 04:16:30", "content": "

Thanks for sharing!

", "rawMarkdown": "Thanks for sharing!", "votes": null }, { "id": "1707033", "postDate": "02/28/2022 04:56:13", "content": "

Welcome @cafelatte1

", "rawMarkdown": "Welcome @cafelatte1", "votes": null }, { "id": "1707034", "postDate": "02/28/2022 04:56:24", "content": "

Welcome @pargo18

", "rawMarkdown": "Welcome @pargo18", "votes": null }, { "id": "1713903", "postDate": "03/06/2022 13:16:57", "content": "

Brilliant!

", "rawMarkdown": "Brilliant!", "votes": null }, { "id": "1714390", "postDate": "03/06/2022 23:57:24", "content": "

Thanks @alexzhuzk

", "rawMarkdown": "Thanks @alexzhuzk", "votes": null }, { "id": "1776135", "postDate": "05/03/2022 17:48:18", "content": "

Thank you, this will be of great help.

", "rawMarkdown": "Thank you, this will be of great help.", "votes": null } ], "comments": [ { "id": 1681672, "author_name": "subhamjain", "author_url": "", "post_date": "02/08/2022 16:13:49", "content": "

Thanks for sharing @odins0n

", "votes": null, "replies": [ { "id": 1681686, "author_name": "odins0n", "author_url": "", "post_date": "02/08/2022 16:26:58", "content": "

Welcome @subhamjain

", "votes": null, "replies": [] } ] }, { "id": 1682322, "author_name": "sanjeevprajapati2", "author_url": "", "post_date": "02/09/2022 04:51:41", "content": "

Thanks a lot for sharing @odins0n
\nI was looking for 256x256. Any chance we can get 224x224 ?

", "votes": null, "replies": [ { "id": 1682576, "author_name": "odins0n", "author_url": "", "post_date": "02/09/2022 08:53:06", "content": "

Welcome @sanjeevprajapati2
\nYou can use this script to generate a dataset of any dimension you want - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256

", "votes": null, "replies": [] } ] }, { "id": 1682424, "author_name": "rahulshastri1", "author_url": "", "post_date": "02/09/2022 06:53:07", "content": "

The dataset size reduced significantly. Thanks for sharing

", "votes": null, "replies": [ { "id": 1682575, "author_name": "odins0n", "author_url": "", "post_date": "02/09/2022 08:52:24", "content": "

Welcome @rahulshastri1

", "votes": null, "replies": [] } ] }, { "id": 1695822, "author_name": "paweljankiewicz", "author_url": "", "post_date": "02/18/2022 11:41:15", "content": "

Thank you so much for this.

", "votes": null, "replies": [ { "id": 1695969, "author_name": "odins0n", "author_url": "", "post_date": "02/18/2022 14:06:56", "content": "

Welcome @paweljankiewicz

", "votes": null, "replies": [] } ] }, { "id": 1698090, "author_name": "init27", "author_url": "", "post_date": "02/20/2022 06:11:10", "content": "

Thanks for creating these @odins0n!

\n

Going through the images, they look like perfectly centered product images.

\n

I suspect we can apply center cropping safely without any issues-maybe an idea for v2 of your datasets? :)

", "votes": null, "replies": [ { "id": 1698155, "author_name": "odins0n", "author_url": "", "post_date": "02/20/2022 07:24:18", "content": "

Welcome @init27
\nThat's a great idea. I will surely try out centre cropping and check for any difference in results.

", "votes": null, "replies": [] } ] }, { "id": 1698673, "author_name": "vad13irt", "author_url": "", "post_date": "02/20/2022 15:13:40", "content": "

Thank you for spending some time doing it for us. I appreciate it, @odins0n!

", "votes": null, "replies": [ { "id": 1698686, "author_name": "odins0n", "author_url": "", "post_date": "02/20/2022 15:23:09", "content": "

Welcome @vad13irt

", "votes": null, "replies": [] } ] }, { "id": 1702148, "author_name": "datascientistfp", "author_url": "", "post_date": "02/23/2022 11:20:56", "content": "

Thank you very much for sharing @odins0n
\nReally helpful! Always good ideas 🧐

", "votes": null, "replies": [ { "id": 1702188, "author_name": "odins0n", "author_url": "", "post_date": "02/23/2022 11:50:12", "content": "

Welcome @datascientistfp

", "votes": null, "replies": [] } ] }, { "id": 1705958, "author_name": "cafelatte1", "author_url": "", "post_date": "02/27/2022 02:45:26", "content": "

WOW Thanks for sharing !

", "votes": null, "replies": [ { "id": 1707033, "author_name": "odins0n", "author_url": "", "post_date": "02/28/2022 04:56:13", "content": "

Welcome @cafelatte1

", "votes": null, "replies": [] } ] }, { "id": 1707015, "author_name": "pargo18", "author_url": "", "post_date": "02/28/2022 04:16:30", "content": "

Thanks for sharing!

", "votes": null, "replies": [ { "id": 1707034, "author_name": "odins0n", "author_url": "", "post_date": "02/28/2022 04:56:24", "content": "

Welcome @pargo18

", "votes": null, "replies": [] } ] }, { "id": 1713903, "author_name": "alexzhuzk", "author_url": "", "post_date": "03/06/2022 13:16:57", "content": "

Brilliant!

", "votes": null, "replies": [ { "id": 1714390, "author_name": "odins0n", "author_url": "", "post_date": "03/06/2022 23:57:24", "content": "

Thanks @alexzhuzk

", "votes": null, "replies": [] } ] }, { "id": 1776135, "author_name": "gkdivya", "author_url": "", "post_date": "05/03/2022 17:48:18", "content": "

Thank you, this will be of great help.

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1681218": "Hello everyone! \n\nI have created 5 different datasets for this competition. I have taken the original images and scaled them down to 5 different lower resolutions (128 x 128, 256 x 256, 384 x 384, 512 x 512 and \n1024 x1024 ). From previous competitions, we have seen that higher resolutions generally have better model performance but these provide a great starting point.\n\nAll 5 datasets also include the `4 tabular CSV` files in each of these datasets for convenience.\n\nThese datasets are also a lot smaller in scale compared to the original dataset which should make it easier to download locally or even use in a Kaggle kernel. \nPlease let me know if you have any feedback ✨\n\nHere are the dataset links:\n\n* (128 x 128 dataset) - https://www.kaggle.com/odins0n/handm-dataset-128x128\n* (256 x 256 dataset) - https://www.kaggle.com/odins0n/hm256x256\n* (384 x 384 dataset) - https://www.kaggle.com/odins0n/hm384x384\n* (512 x 512 dataset) - https://www.kaggle.com/odins0n/hm512x512\n* (1024 x 1024 dataset) - https://www.kaggle.com/odins0n/hm1024x1024\n\nIf you want are interested in knowing how the data was compressed and which interpolation method was used, check out this notebook - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256", "1681672": "Thanks for sharing @odins0n", "1681686": "Welcome @subhamjain", "1682322": "Thanks a lot for sharing @odins0n \nI was looking for 256x256. Any chance we can get 224x224 ?", "1682424": "The dataset size reduced significantly. Thanks for sharing", "1682575": "Welcome @rahulshastri1", "1682576": "Welcome @sanjeevprajapati2 \nYou can use this script to generate a dataset of any dimension you want - https://www.kaggle.com/odins0n/h-m-image-resizer-dataset-256x256", "1695822": "Thank you so much for this.", "1695969": "Welcome @paweljankiewicz", "1698090": "Thanks for creating these @odins0n! \n\nGoing through the images, they look like perfectly centered product images. \n\nI suspect we can apply center cropping safely without any issues-maybe an idea for v2 of your datasets? :)", "1698155": "Welcome @init27 \nThat's a great idea. I will surely try out centre cropping and check for any difference in results.", "1698673": "Thank you for spending some time doing it for us. I appreciate it, @odins0n!", "1698686": "Welcome @vad13irt", "1702148": "Thank you very much for sharing @odins0n \nReally helpful! Always good ideas 🧐", "1702188": "Welcome @datascientistfp", "1705958": "WOW Thanks for sharing !", "1707015": "Thanks for sharing!", "1707033": "Welcome @cafelatte1", "1707034": "Welcome @pargo18", "1713903": "Brilliant!", "1714390": "Thanks @alexzhuzk", "1776135": "Thank you, this will be of great help." }, "source": "meta" }