{
  "id": 198050,
  "title": "Fast and memory-saving Image Loading using Pyvips",
  "url": "/competitions/hubmap-kidney-segmentation/discussion/198050",
  "author_name": "",
  "post_date": "2020-11-19T14:20:32.870241500Z",
  "votes": 32,
  "comment_count": 4,
  "views": 0,
  "content": "<p>The images in this competition are very large and consume a lot of memory when loaded in their entirety.<br>\nHere, you can use pyvips to read only some areas. Please refer to the following notebook.<br>\n<a href=\"https://www.kaggle.com/hirune924/fast-image-region-loading-using-pyvips\" target=\"_blank\">https://www.kaggle.com/hirune924/fast-image-region-loading-using-pyvips</a><br>\nIt's difficult to install this library without an internet connection, so I'd like you to add this library to your kaggle environment. What do you think?</p>",
  "messages": [
    {
      "id": "1083907",
      "postDate": "11/19/2020 14:20:32",
      "content": "<p>The images in this competition are very large and consume a lot of memory when loaded in their entirety.<br>\nHere, you can use pyvips to read only some areas. Please refer to the following notebook.<br>\n<a href=\"https://www.kaggle.com/hirune924/fast-image-region-loading-using-pyvips\" target=\"_blank\">https://www.kaggle.com/hirune924/fast-image-region-loading-using-pyvips</a><br>\nIt's difficult to install this library without an internet connection, so I'd like you to add this library to your kaggle environment. What do you think?</p>",
      "rawMarkdown": "The images in this competition are very large and consume a lot of memory when loaded in their entirety.\nHere, you can use pyvips to read only some areas. Please refer to the following notebook.\nhttps://www.kaggle.com/hirune924/fast-image-region-loading-using-pyvips\n\nIt's difficult to install this library without an internet connection, so I'd like you to add this library to your kaggle environment. What do you think?",
      "votes": null
    },
    {
      "id": "1085114",
      "postDate": "11/20/2020 17:17:59",
      "content": "<p>This is amazing, thanks for figuring out how to install it inside Kaggle notebooks! This looks super useful for both loading in memory and directly saving each patch into png files</p>",
      "rawMarkdown": "This is amazing, thanks for figuring out how to install it inside Kaggle notebooks! This looks super useful for both loading in memory and directly saving each patch into png files",
      "votes": null
    },
    {
      "id": "1085237",
      "postDate": "11/20/2020 19:19:04",
      "content": "<p><a href=\"https://www.kaggle.com/hirune924\" target=\"_blank\">@hirune924</a> Thanks for the super helpful to save memory and load image faster .</p>",
      "rawMarkdown": "hirune924 Thanks for the super helpful to save memory and load image faster .",
      "votes": null
    },
    {
      "id": "1090031",
      "postDate": "11/25/2020 02:16:00",
      "content": "<p>thank you for cool code !<br>\nit was really effective way for me to avoid oom😉<br>\nnext problem is internet isn't allowed for submission kernel (can't directly use !pip install pyvips)…😇</p>",
      "rawMarkdown": "thank you for cool code !\nit was really effective way for me to avoid oom😉\nnext problem is internet isn't allowed for submission kernel (can't directly use !pip install pyvips)...😇",
      "votes": null
    },
    {
      "id": "1090103",
      "postDate": "11/25/2020 04:44:46",
      "content": "<p>Nice Work! It's cool !</p>",
      "rawMarkdown": "Nice Work! It's cool !",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1085114,
      "author_name": "xhlulu",
      "author_url": "",
      "post_date": "11/20/2020 17:17:59",
      "content": "<p>This is amazing, thanks for figuring out how to install it inside Kaggle notebooks! This looks super useful for both loading in memory and directly saving each patch into png files</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1085237,
      "author_name": "usharengaraju",
      "author_url": "",
      "post_date": "11/20/2020 19:19:04",
      "content": "<p><a href=\"https://www.kaggle.com/hirune924\" target=\"_blank\">@hirune924</a> Thanks for the super helpful to save memory and load image faster .</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1090031,
      "author_name": "ryunosukeishizaki",
      "author_url": "",
      "post_date": "11/25/2020 02:16:00",
      "content": "<p>thank you for cool code !<br>\nit was really effective way for me to avoid oom😉<br>\nnext problem is internet isn't allowed for submission kernel (can't directly use !pip install pyvips)…😇</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1090103,
      "author_name": "kaushlendrat",
      "author_url": "",
      "post_date": "11/25/2020 04:44:46",
      "content": "<p>Nice Work! It's cool !</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1083907": "The images in this competition are very large and consume a lot of memory when loaded in their entirety.\nHere, you can use pyvips to read only some areas. Please refer to the following notebook.\nhttps://www.kaggle.com/hirune924/fast-image-region-loading-using-pyvips\n\nIt's difficult to install this library without an internet connection, so I'd like you to add this library to your kaggle environment. What do you think?",
    "1085114": "This is amazing, thanks for figuring out how to install it inside Kaggle notebooks! This looks super useful for both loading in memory and directly saving each patch into png files",
    "1085237": "hirune924 Thanks for the super helpful to save memory and load image faster .",
    "1090031": "thank you for cool code !\nit was really effective way for me to avoid oom😉\nnext problem is internet isn't allowed for submission kernel (can't directly use !pip install pyvips)...😇",
    "1090103": "Nice Work! It's cool !"
  },
  "source": "meta"
}