{
  "id": 111499,
  "title": "How apply random crop with mask pixel information?",
  "url": "/competitions/understanding_cloud_organization/discussion/111499",
  "author_name": "",
  "post_date": "2019-10-06T10:41:47.633145400Z",
  "votes": 2,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I'm looking for method that perform random crop  to both image and mask pixel. </p>\n\n<p>Is there any solution for this?</p>",
  "messages": [
    {
      "id": "642587",
      "postDate": "10/06/2019 10:41:47",
      "content": "<p>I'm looking for method that perform random crop  to both image and mask pixel. </p>\n\n<p>Is there any solution for this?</p>",
      "rawMarkdown": "I'm looking for method that perform random crop  to both image and mask pixel. \n\nIs there any solution for this?",
      "votes": null
    },
    {
      "id": "642600",
      "postDate": "10/06/2019 11:19:16",
      "content": "<p>Try albumentations: <a href=\"https://albumentations.readthedocs.io/en/latest/api/augmentations.html#albumentations.augmentations.transforms.RandomCrop\">https://albumentations.readthedocs.io/en/latest/api/augmentations.html#albumentations.augmentations.transforms.RandomCrop</a></p>",
      "rawMarkdown": "Try albumentations: https://albumentations.readthedocs.io/en/latest/api/augmentations.html#albumentations.augmentations.transforms.RandomCrop",
      "votes": null
    },
    {
      "id": "642617",
      "postDate": "10/06/2019 11:52:30",
      "content": "<p>Thank you\nI was trying to apply transformations in native pytorch, I followed this tutorial: <a href=\"https://pytorch.org/tutorials/beginner/data_loading_tutorial.html\">https://pytorch.org/tutorials/beginner/data_loading_tutorial.html</a>\nBut it seems like they have written their transformations from scratch which is pretty tiresome.\nI saw other kernels here, but all of them seem to use albumations. I guess we have no other choice. </p>",
      "rawMarkdown": "Thank you\nI was trying to apply transformations in native pytorch, I followed this tutorial: https://pytorch.org/tutorials/beginner/data_loading_tutorial.html\nBut it seems like they have written their transformations from scratch which is pretty tiresome.\nI saw other kernels here, but all of them seem to use albumations. I guess we have no other choice.",
      "votes": null
    },
    {
      "id": "642656",
      "postDate": "10/06/2019 12:47:58",
      "content": "<p>There is a reason everyone uses <code>albumentations</code> library; it's really convenient and gives you a lot of options to choose from for your augmentations </p>",
      "rawMarkdown": "There is a reason everyone uses `albumentations` library; it's really convenient and gives you a lot of options to choose from for your augmentations",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 642600,
      "author_name": "artgor",
      "author_url": "",
      "post_date": "10/06/2019 11:19:16",
      "content": "<p>Try albumentations: <a href=\"https://albumentations.readthedocs.io/en/latest/api/augmentations.html#albumentations.augmentations.transforms.RandomCrop\">https://albumentations.readthedocs.io/en/latest/api/augmentations.html#albumentations.augmentations.transforms.RandomCrop</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 642617,
          "author_name": "timetraveller98",
          "author_url": "",
          "post_date": "10/06/2019 11:52:30",
          "content": "<p>Thank you\nI was trying to apply transformations in native pytorch, I followed this tutorial: <a href=\"https://pytorch.org/tutorials/beginner/data_loading_tutorial.html\">https://pytorch.org/tutorials/beginner/data_loading_tutorial.html</a>\nBut it seems like they have written their transformations from scratch which is pretty tiresome.\nI saw other kernels here, but all of them seem to use albumations. I guess we have no other choice. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 642656,
          "author_name": "bibek777",
          "author_url": "",
          "post_date": "10/06/2019 12:47:58",
          "content": "<p>There is a reason everyone uses <code>albumentations</code> library; it's really convenient and gives you a lot of options to choose from for your augmentations </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "642587": "I'm looking for method that perform random crop  to both image and mask pixel. \n\nIs there any solution for this?",
    "642600": "Try albumentations: https://albumentations.readthedocs.io/en/latest/api/augmentations.html#albumentations.augmentations.transforms.RandomCrop",
    "642617": "Thank you\nI was trying to apply transformations in native pytorch, I followed this tutorial: https://pytorch.org/tutorials/beginner/data_loading_tutorial.html\nBut it seems like they have written their transformations from scratch which is pretty tiresome.\nI saw other kernels here, but all of them seem to use albumations. I guess we have no other choice.",
    "642656": "There is a reason everyone uses `albumentations` library; it's really convenient and gives you a lot of options to choose from for your augmentations"
  },
  "source": "meta"
}