{
  "id": 66269,
  "title": "Questions about the labels",
  "url": "/competitions/inclusive-images-challenge/discussion/66269",
  "author_name": "",
  "post_date": "2018-09-19T20:43:00.489298800Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi, I have read the FAQ but it's still a bit unclear.</p>\n\n<p>The labels in the output should be among the labels in classes-trainable.csv, right?</p>\n\n<p>But there are some labels in train_human_labels.csv which are not contained in classes-trainable.csv. Should I ignore them?</p>\n\n<p>Also, train_machine_labels.csv is given but I cannot find any information about the model. Is it given to use as ground truth data? or is it just for reference?</p>\n\n<p>Thanks!</p>",
  "messages": [
    {
      "id": "390185",
      "postDate": "09/19/2018 20:43:00",
      "content": "<p>Hi, I have read the FAQ but it's still a bit unclear.</p>\n\n<p>The labels in the output should be among the labels in classes-trainable.csv, right?</p>\n\n<p>But there are some labels in train_human_labels.csv which are not contained in classes-trainable.csv. Should I ignore them?</p>\n\n<p>Also, train_machine_labels.csv is given but I cannot find any information about the model. Is it given to use as ground truth data? or is it just for reference?</p>\n\n<p>Thanks!</p>",
      "rawMarkdown": "Hi, I have read the FAQ but it's still a bit unclear.\n\nThe labels in the output should be among the labels in classes-trainable.csv, right?\n\nBut there are some labels in train_human_labels.csv which are not contained in classes-trainable.csv. Should I ignore them?\n\nAlso, train_machine_labels.csv is given but I cannot find any information about the model. Is it given to use as ground truth data? or is it just for reference?\n\nThanks!",
      "votes": null
    },
    {
      "id": "390603",
      "postDate": "09/20/2018 13:58:41",
      "content": "<p>Hi Syd,\nYes you need to only use the classes in the classes-trainable.csv, and ignore the rest. You can choose to use the human and machine labels as you see fit.</p>",
      "rawMarkdown": "Hi Syd,\nYes you need to only use the classes in the classes-trainable.csv, and ignore the rest. You can choose to use the human and machine labels as you see fit.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 390603,
      "author_name": "pbaljeka",
      "author_url": "",
      "post_date": "09/20/2018 13:58:41",
      "content": "<p>Hi Syd,\nYes you need to only use the classes in the classes-trainable.csv, and ignore the rest. You can choose to use the human and machine labels as you see fit.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "390185": "Hi, I have read the FAQ but it's still a bit unclear.\n\nThe labels in the output should be among the labels in classes-trainable.csv, right?\n\nBut there are some labels in train_human_labels.csv which are not contained in classes-trainable.csv. Should I ignore them?\n\nAlso, train_machine_labels.csv is given but I cannot find any information about the model. Is it given to use as ground truth data? or is it just for reference?\n\nThanks!",
    "390603": "Hi Syd,\nYes you need to only use the classes in the classes-trainable.csv, and ignore the rest. You can choose to use the human and machine labels as you see fit."
  },
  "source": "meta"
}