{
  "id": 130440,
  "title": "What do we need to predict?",
  "url": "/competitions/deepfake-detection-challenge/discussion/130440",
  "author_name": "",
  "post_date": "2020-02-14T06:06:43.171163500Z",
  "votes": null,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Greetings, </p>\n\n<p>I am starting with the challenge and was finding it difficult to understand the following:</p>\n\n<ol>\n<li><p>What do we need to predict?</p></li>\n<li><p>What is our X and what should be our Y</p></li>\n<li><p>Is it a classification challenge or a prediction challenge?</p></li>\n</ol>\n\n<p>Kindly let me know.</p>\n\n<p>Note: If you find the questions naive kindly accept my apologies I am not that \"hifi\" :-)</p>\n\n<p>Best regards</p>",
  "messages": [
    {
      "id": "745729",
      "postDate": "02/14/2020 06:06:43",
      "content": "<p>Greetings, </p>\n\n<p>I am starting with the challenge and was finding it difficult to understand the following:</p>\n\n<ol>\n<li><p>What do we need to predict?</p></li>\n<li><p>What is our X and what should be our Y</p></li>\n<li><p>Is it a classification challenge or a prediction challenge?</p></li>\n</ol>\n\n<p>Kindly let me know.</p>\n\n<p>Note: If you find the questions naive kindly accept my apologies I am not that \"hifi\" :-)</p>\n\n<p>Best regards</p>",
      "rawMarkdown": "Greetings, \n\n\nI am starting with the challenge and was finding it difficult to understand the following:\n\n1. What do we need to predict?\n\n2. What is our X and what should be our Y\n\n3. Is it a classification challenge or a prediction challenge?\n\nKindly let me know.\n\nNote: If you find the questions naive kindly accept my apologies I am not that \"hifi\" :-)\n\nBest regards",
      "votes": null
    },
    {
      "id": "745735",
      "postDate": "02/14/2020 06:16:12",
      "content": "<ol>\n<li>Real versus fake</li>\n<li>X = video, y = prediction </li>\n<li>Binary classification</li>\n</ol>",
      "rawMarkdown": "1. Real versus fake\n2. X = video, y = prediction \n3. Binary classification",
      "votes": null
    },
    {
      "id": "745782",
      "postDate": "02/14/2020 07:42:48",
      "content": "<p>Thanks Nick.\nI am assuming that 0 is a fake and 1 is real for prediction?. Is there a sample file submission example file?</p>",
      "rawMarkdown": "Thanks Nick.\nI am assuming that 0 is a fake and 1 is real for prediction?. Is there a sample file submission example file?",
      "votes": null
    },
    {
      "id": "745884",
      "postDate": "02/14/2020 10:12:20",
      "content": "<p>It's the other way around. 1 is the label for fake and 0 is the label for real. Check out the \"Evaluation\" section for an example of the submission file structure or some public kernels.</p>",
      "rawMarkdown": "It's the other way around. 1 is the label for fake and 0 is the label for real. Check out the \"Evaluation\" section for an example of the submission file structure or some public kernels.",
      "votes": null
    },
    {
      "id": "746624",
      "postDate": "02/15/2020 09:29:27",
      "content": "<p><a href=\"/catochris\">@catochris</a>  my label process takes 0 as fake and 1 as real. my model gives two outputs ,so i take class[:,0]  as submission prob. is that fine right ?</p>",
      "rawMarkdown": "catochris  my label process takes 0 as fake and 1 as real. my model gives two outputs ,so i take class[:,0]  as submission prob. is that fine right ?",
      "votes": null
    },
    {
      "id": "746642",
      "postDate": "02/15/2020 09:58:07",
      "content": "<p>Should be fine if the probability at class[:,0] is the probability that the video is fake.</p>",
      "rawMarkdown": "Should be fine if the probability at class[:,0] is the probability that the video is fake.",
      "votes": null
    },
    {
      "id": "747671",
      "postDate": "02/16/2020 17:58:01",
      "content": "<p>I recommend you to only do 1 output.</p>",
      "rawMarkdown": "I recommend you to only do 1 output.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 745735,
      "author_name": "sciarrilli",
      "author_url": "",
      "post_date": "02/14/2020 06:16:12",
      "content": "<ol>\n<li>Real versus fake</li>\n<li>X = video, y = prediction </li>\n<li>Binary classification</li>\n</ol>",
      "votes": null,
      "replies": []
    },
    {
      "id": 745782,
      "author_name": "ds123456",
      "author_url": "",
      "post_date": "02/14/2020 07:42:48",
      "content": "<p>Thanks Nick.\nI am assuming that 0 is a fake and 1 is real for prediction?. Is there a sample file submission example file?</p>",
      "votes": null,
      "replies": [
        {
          "id": 745884,
          "author_name": "catochris",
          "author_url": "",
          "post_date": "02/14/2020 10:12:20",
          "content": "<p>It's the other way around. 1 is the label for fake and 0 is the label for real. Check out the \"Evaluation\" section for an example of the submission file structure or some public kernels.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 746624,
      "author_name": "jaideepvalani",
      "author_url": "",
      "post_date": "02/15/2020 09:29:27",
      "content": "<p><a href=\"/catochris\">@catochris</a>  my label process takes 0 as fake and 1 as real. my model gives two outputs ,so i take class[:,0]  as submission prob. is that fine right ?</p>",
      "votes": null,
      "replies": [
        {
          "id": 746642,
          "author_name": "catochris",
          "author_url": "",
          "post_date": "02/15/2020 09:58:07",
          "content": "<p>Should be fine if the probability at class[:,0] is the probability that the video is fake.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 747671,
          "author_name": "unkownhihi",
          "author_url": "",
          "post_date": "02/16/2020 17:58:01",
          "content": "<p>I recommend you to only do 1 output.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "745729": "Greetings, \n\n\nI am starting with the challenge and was finding it difficult to understand the following:\n\n1. What do we need to predict?\n\n2. What is our X and what should be our Y\n\n3. Is it a classification challenge or a prediction challenge?\n\nKindly let me know.\n\nNote: If you find the questions naive kindly accept my apologies I am not that \"hifi\" :-)\n\nBest regards",
    "745735": "1. Real versus fake\n2. X = video, y = prediction \n3. Binary classification",
    "745782": "Thanks Nick.\nI am assuming that 0 is a fake and 1 is real for prediction?. Is there a sample file submission example file?",
    "745884": "It's the other way around. 1 is the label for fake and 0 is the label for real. Check out the \"Evaluation\" section for an example of the submission file structure or some public kernels.",
    "746624": "catochris  my label process takes 0 as fake and 1 as real. my model gives two outputs ,so i take class[:,0]  as submission prob. is that fine right ?",
    "746642": "Should be fine if the probability at class[:,0] is the probability that the video is fake.",
    "747671": "I recommend you to only do 1 output."
  },
  "source": "meta"
}