{
  "id": 155128,
  "title": "Downloading dataset",
  "url": "/competitions/open-images-object-detection-rvc-2020/discussion/155128",
  "author_name": "Adrien I",
  "post_date": "2020-05-31T13:08:25.923000",
  "votes": 0,
  "comment_count": 11,
  "views": 0,
  "content": "<p>Hello, I don't really understand what I need to download to do this competition.\nIf anybody can help me?</p>",
  "messages": [
    {
      "id": 868764,
      "postDate": "2020-05-31T13:21:45.270Z",
      "content": "<p>There are 99999 images in the test folder. I think you should download them and start the competition.</p>",
      "rawMarkdown": "There are 99999 images in the test folder. I think you should download them and start the competition.",
      "votes": 1,
      "replies": [
        {
          "id": 959501,
          "postDate": "2020-08-05T16:19:18.747Z",
          "content": "<p>That is the test folder, the training set is in another competition. See link in a the 'Welcome' discussion in this competition.</p>",
          "rawMarkdown": "That is the test folder, the training set is in another competition. See link in a the 'Welcome' discussion in this competition."
        }
      ]
    },
    {
      "id": 873068,
      "postDate": "2020-06-03T18:40:24.833Z",
      "content": "<p>It's a bit of a scavenger hunt. The data page for this competition has a link to the open images site which then has links to what you need for training and/or evaluation. You need at a minimum the annotation files with boxes and labels, and the images. There are also metadata files that can help. </p>\n\n<p>I show how to get started with files in this <a href=\"https://www.kaggle.com/jpmiller/open-images-eda\">notebook</a>. The segment masks are for the other challenge so no need to worry about those.</p>",
      "rawMarkdown": "It's a bit of a scavenger hunt. The data page for this competition has a link to the open images site which then has links to what you need for training and/or evaluation. You need at a minimum the annotation files with boxes and labels, and the images. There are also metadata files that can help. \n\nI show how to get started with files in this [notebook](https://www.kaggle.com/jpmiller/open-images-eda). The segment masks are for the other challenge so no need to worry about those.",
      "votes": 2,
      "replies": [
        {
          "id": 951591,
          "postDate": "2020-07-30T08:39:46.297Z",
          "content": "<p>Hallo, there is a problem, it seems that I need a AWS or a Google account in order to download the images, so I cannot do it. Have you any idea where can I download the images with the bounding boxes  and labels ?</p>",
          "rawMarkdown": "Hallo, there is a problem, it seems that I need a AWS or a Google account in order to download the images, so I cannot do it. Have you any idea where can I download the images with the bounding boxes  and labels ?"
        },
        {
          "id": 951986,
          "postDate": "2020-07-30T14:35:47.850Z",
          "content": "<p>Maybe try <a href=\"https://storage.googleapis.com/openimages/web/download.html\">https://storage.googleapis.com/openimages/web/download.html</a></p>",
          "rawMarkdown": "Maybe try https://storage.googleapis.com/openimages/web/download.html\n"
        },
        {
          "id": 959450,
          "postDate": "2020-08-05T15:36:04.913Z",
          "rawMarkdown": "",
          "isDeleted": true
        },
        {
          "id": 959497,
          "postDate": "2020-08-05T16:17:30.047Z",
          "rawMarkdown": "",
          "isDeleted": true
        }
      ]
    },
    {
      "id": 959565,
      "postDate": "2020-08-05T17:39:13.643Z",
      "content": "<p>Training set <a href=\"https://www.kaggle.com/c/open-images-object-detection-rvc-2020/data?select=test\">here</a>, according to answer in <a href=\"https://www.kaggle.com/c/open-images-object-detection-rvc-2020/discussion/152429\">this discussion</a>.</p>",
      "rawMarkdown": "Training set [here](https://www.kaggle.com/c/open-images-object-detection-rvc-2020/data?select=test), according to answer in [this discussion](https://www.kaggle.com/c/open-images-object-detection-rvc-2020/discussion/152429)."
    },
    {
      "id": 868803,
      "postDate": "2020-05-31T14:01:32.010Z",
      "content": "<p>ok</p>",
      "rawMarkdown": "ok"
    },
    {
      "id": 868779,
      "postDate": "2020-05-31T13:41:27.703Z",
      "content": "<p>Ok, but I want to download the train and validation dataset</p>",
      "rawMarkdown": "Ok, but I want to download the train and validation dataset",
      "replies": [
        {
          "id": 868799,
          "postDate": "2020-05-31T13:59:36.643Z",
          "content": "<p>you can use train_test_split() twice. After first split you are gonna obtain train and test parts. After split the \"train\" part with using same method. You are gonna obtain train and validation parts.</p>",
          "rawMarkdown": "you can use train_test_split() twice. After first split you are gonna obtain train and test parts. After split the \"train\" part with using same method. You are gonna obtain train and validation parts."
        }
      ]
    },
    {
      "id": 868748,
      "postDate": "2020-05-31T13:08:25.923Z",
      "content": "<p>Hello, I don't really understand what I need to download to do this competition.\nIf anybody can help me?</p>",
      "rawMarkdown": "Hello, I don't really understand what I need to download to do this competition.\nIf anybody can help me?"
    }
  ],
  "comments": [
    {
      "id": 868764,
      "author_name": "Cihan Senol",
      "author_url": "",
      "post_date": "2020-05-31T13:21:45.270000",
      "content": "<p>There are 99999 images in the test folder. I think you should download them and start the competition.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 959501,
          "author_name": "Catadanna",
          "author_url": "",
          "post_date": "2020-08-05T16:19:18.747000",
          "content": "<p>That is the test folder, the training set is in another competition. See link in a the 'Welcome' discussion in this competition.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 873068,
      "author_name": "JohnM",
      "author_url": "",
      "post_date": "2020-06-03T18:40:24.833000",
      "content": "<p>It's a bit of a scavenger hunt. The data page for this competition has a link to the open images site which then has links to what you need for training and/or evaluation. You need at a minimum the annotation files with boxes and labels, and the images. There are also metadata files that can help. </p>\n\n<p>I show how to get started with files in this <a href=\"https://www.kaggle.com/jpmiller/open-images-eda\">notebook</a>. The segment masks are for the other challenge so no need to worry about those.</p>",
      "votes": 2,
      "replies": [
        {
          "id": 951591,
          "author_name": "Catadanna",
          "author_url": "",
          "post_date": "2020-07-30T08:39:46.297000",
          "content": "<p>Hallo, there is a problem, it seems that I need a AWS or a Google account in order to download the images, so I cannot do it. Have you any idea where can I download the images with the bounding boxes  and labels ?</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 951986,
          "author_name": "JohnM",
          "author_url": "",
          "post_date": "2020-07-30T14:35:47.850000",
          "content": "<p>Maybe try <a href=\"https://storage.googleapis.com/openimages/web/download.html\">https://storage.googleapis.com/openimages/web/download.html</a></p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 959450,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-08-05T15:36:04.913000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 959497,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-08-05T16:17:30.047000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 959565,
      "author_name": "Catadanna",
      "author_url": "",
      "post_date": "2020-08-05T17:39:13.643000",
      "content": "<p>Training set <a href=\"https://www.kaggle.com/c/open-images-object-detection-rvc-2020/data?select=test\">here</a>, according to answer in <a href=\"https://www.kaggle.com/c/open-images-object-detection-rvc-2020/discussion/152429\">this discussion</a>.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 868803,
      "author_name": "Adrien I",
      "author_url": "",
      "post_date": "2020-05-31T14:01:32.010000",
      "content": "<p>ok</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 868779,
      "author_name": "Adrien I",
      "author_url": "",
      "post_date": "2020-05-31T13:41:27.703000",
      "content": "<p>Ok, but I want to download the train and validation dataset</p>",
      "votes": 0,
      "replies": [
        {
          "id": 868799,
          "author_name": "Cihan Senol",
          "author_url": "",
          "post_date": "2020-05-31T13:59:36.643000",
          "content": "<p>you can use train_test_split() twice. After first split you are gonna obtain train and test parts. After split the \"train\" part with using same method. You are gonna obtain train and validation parts.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "868764": "There are 99999 images in the test folder. I think you should download them and start the competition.",
    "873068": "It's a bit of a scavenger hunt. The data page for this competition has a link to the open images site which then has links to what you need for training and/or evaluation. You need at a minimum the annotation files with boxes and labels, and the images. There are also metadata files that can help. \n\nI show how to get started with files in this [notebook](https://www.kaggle.com/jpmiller/open-images-eda). The segment masks are for the other challenge so no need to worry about those.",
    "959565": "Training set [here](https://www.kaggle.com/c/open-images-object-detection-rvc-2020/data?select=test), according to answer in [this discussion](https://www.kaggle.com/c/open-images-object-detection-rvc-2020/discussion/152429).",
    "868803": "ok",
    "868779": "Ok, but I want to download the train and validation dataset",
    "868748": "Hello, I don't really understand what I need to download to do this competition.\nIf anybody can help me?"
  }
}