{
  "id": 41374,
  "title": "Some help need to verify quality of my train.bson",
  "url": "/competitions/cdiscount-image-classification-challenge/discussion/41374",
  "author_name": "",
  "post_date": "2017-10-17T02:41:11.542340900Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Can someone look at image in train.bson for product id 10076032?\nIt has 4 images.\nDoes it look like it belongs to category ID \n1000009283,AUTO - MOTO,OUTILLAGE,TRACEUR DE FUITE - KIT TRACEUR DE FUITE ???\nI am trying to make sure that my downloaded train.bson is not corrupted and nor is my image offset wrong.\nHelp from any quarter is appreciated !!!\nMerci'</p>",
  "messages": [
    {
      "id": "232184",
      "postDate": "10/17/2017 02:41:11",
      "content": "<p>Can someone look at image in train.bson for product id 10076032?\nIt has 4 images.\nDoes it look like it belongs to category ID \n1000009283,AUTO - MOTO,OUTILLAGE,TRACEUR DE FUITE - KIT TRACEUR DE FUITE ???\nI am trying to make sure that my downloaded train.bson is not corrupted and nor is my image offset wrong.\nHelp from any quarter is appreciated !!!\nMerci'</p>",
      "rawMarkdown": "Can someone look at image in train.bson for product id 10076032?\nIt has 4 images.\nDoes it look like it belongs to category ID \n1000009283,AUTO - MOTO,OUTILLAGE,TRACEUR DE FUITE - KIT TRACEUR DE FUITE ???\nI am trying to make sure that my downloaded train.bson is not corrupted and nor is my image offset wrong.\nHelp from any quarter is appreciated !!!\nMerci'",
      "votes": null
    },
    {
      "id": "232232",
      "postDate": "10/17/2017 06:45:53",
      "content": "<p>There is no perfect-labeled dataset with &gt;12M images.</p>\n\n<p>Anyway you can validate your files like that: <a href=\"https://www.kaggle.com/blazeka/validate-download-with-sha256-hash\">https://www.kaggle.com/blazeka/validate-download-with-sha256-hash</a>.</p>",
      "rawMarkdown": "There is no perfect-labeled dataset with &gt;12M images.\n\nAnyway you can validate your files like that: https://www.kaggle.com/blazeka/validate-download-with-sha256-hash.",
      "votes": null
    },
    {
      "id": "232259",
      "postDate": "10/17/2017 08:12:00",
      "content": "<p>From a simple inspection of the dataset I've seen quite a few images that do not appear to make sense for that category, or that are completely empty, etc. There are mistakes in any dataset.</p>",
      "rawMarkdown": "From a simple inspection of the dataset I've seen quite a few images that do not appear to make sense for that category, or that are completely empty, etc. There are mistakes in any dataset.",
      "votes": null
    },
    {
      "id": "232371",
      "postDate": "10/17/2017 12:14:41",
      "content": "<p>@HumanAnalog and @EvgenyNizhibitsky,\nThank you both</p>",
      "rawMarkdown": "HumanAnalog and @EvgenyNizhibitsky,\nThank you both",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 232232,
      "author_name": "nizhibitsky",
      "author_url": "",
      "post_date": "10/17/2017 06:45:53",
      "content": "<p>There is no perfect-labeled dataset with &gt;12M images.</p>\n\n<p>Anyway you can validate your files like that: <a href=\"https://www.kaggle.com/blazeka/validate-download-with-sha256-hash\">https://www.kaggle.com/blazeka/validate-download-with-sha256-hash</a>.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 232259,
      "author_name": "humananalog",
      "author_url": "",
      "post_date": "10/17/2017 08:12:00",
      "content": "<p>From a simple inspection of the dataset I've seen quite a few images that do not appear to make sense for that category, or that are completely empty, etc. There are mistakes in any dataset.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 232371,
      "author_name": "shirishr",
      "author_url": "",
      "post_date": "10/17/2017 12:14:41",
      "content": "<p>@HumanAnalog and @EvgenyNizhibitsky,\nThank you both</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "232184": "Can someone look at image in train.bson for product id 10076032?\nIt has 4 images.\nDoes it look like it belongs to category ID \n1000009283,AUTO - MOTO,OUTILLAGE,TRACEUR DE FUITE - KIT TRACEUR DE FUITE ???\nI am trying to make sure that my downloaded train.bson is not corrupted and nor is my image offset wrong.\nHelp from any quarter is appreciated !!!\nMerci'",
    "232232": "There is no perfect-labeled dataset with &gt;12M images.\n\nAnyway you can validate your files like that: https://www.kaggle.com/blazeka/validate-download-with-sha256-hash.",
    "232259": "From a simple inspection of the dataset I've seen quite a few images that do not appear to make sense for that category, or that are completely empty, etc. There are mistakes in any dataset.",
    "232371": "HumanAnalog and @EvgenyNizhibitsky,\nThank you both"
  },
  "source": "meta"
}