{
  "id": 140281,
  "title": "Request to ground truth answer of test videos",
  "url": "/competitions/deepfake-detection-challenge/discussion/140281",
  "author_name": "",
  "post_date": "2020-04-01T06:36:11.916437800Z",
  "votes": 1,
  "comment_count": 6,
  "views": 0,
  "content": "<p>For further research of deepfake detection, we make request to ground truth answer of test videos from deepfake detection challenge and evaluate performance of new models in the future.</p>",
  "messages": [
    {
      "id": "793612",
      "postDate": "04/01/2020 06:36:11",
      "content": "<p>For further research of deepfake detection, we make request to ground truth answer of test videos from deepfake detection challenge and evaluate performance of new models in the future.</p>",
      "rawMarkdown": "For further research of deepfake detection, we make request to ground truth answer of test videos from deepfake detection challenge and evaluate performance of new models in the future.",
      "votes": null
    },
    {
      "id": "793635",
      "postDate": "04/01/2020 07:06:09",
      "content": "<p>Do you mean the 400 videos? They are from the training set so you have the gt. </p>",
      "rawMarkdown": "Do you mean the 400 videos? They are from the training set so you have the gt.",
      "votes": null
    },
    {
      "id": "793726",
      "postDate": "04/01/2020 08:26:11",
      "content": "<p>@Moshel, by \"from the training set\", do you mean the 500G-training set? or the small sample training set \"input\\deepfake-detection-challenge\\train_sample_videos\" ? Thanks</p>",
      "rawMarkdown": "Moshel, by \"from the training set\", do you mean the 500G-training set? or the small sample training set \"input\\deepfake-detection-challenge\\train_sample_videos\" ? Thanks",
      "votes": null
    },
    {
      "id": "793823",
      "postDate": "04/01/2020 10:14:05",
      "content": "<p>From the 500gb. You can find the Metadata files is some datasets. Look at the public kernels. </p>",
      "rawMarkdown": "From the 500gb. You can find the Metadata files is some datasets. Look at the public kernels.",
      "votes": null
    },
    {
      "id": "794727",
      "postDate": "04/02/2020 03:28:57",
      "content": "<p>Thank you very much! I mean the 400 videos from input/deepfake-detection-challenge/test_videos. We predict these videos and finish the submission.csv.</p>",
      "rawMarkdown": "Thank you very much! I mean the 400 videos from input/deepfake-detection-challenge/test_videos. We predict these videos and finish the submission.csv.",
      "votes": null
    },
    {
      "id": "794741",
      "postDate": "04/02/2020 03:48:13",
      "content": "<p>They are videos from the big training set, from folders 0 and 1 iirc. You can get the gt from the training metadata</p>",
      "rawMarkdown": "They are videos from the big training set, from folders 0 and 1 iirc. You can get the gt from the training metadata",
      "votes": null
    },
    {
      "id": "794748",
      "postDate": "04/02/2020 04:00:09",
      "content": "<p>OK. I check the big training set. Thank you!</p>",
      "rawMarkdown": "OK. I check the big training set. Thank you!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 793635,
      "author_name": "moshel",
      "author_url": "",
      "post_date": "04/01/2020 07:06:09",
      "content": "<p>Do you mean the 400 videos? They are from the training set so you have the gt. </p>",
      "votes": null,
      "replies": [
        {
          "id": 794727,
          "author_name": "zimingyang",
          "author_url": "",
          "post_date": "04/02/2020 03:28:57",
          "content": "<p>Thank you very much! I mean the 400 videos from input/deepfake-detection-challenge/test_videos. We predict these videos and finish the submission.csv.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 794741,
          "author_name": "moshel",
          "author_url": "",
          "post_date": "04/02/2020 03:48:13",
          "content": "<p>They are videos from the big training set, from folders 0 and 1 iirc. You can get the gt from the training metadata</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 794748,
          "author_name": "zimingyang",
          "author_url": "",
          "post_date": "04/02/2020 04:00:09",
          "content": "<p>OK. I check the big training set. Thank you!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 793726,
      "author_name": "minghuang",
      "author_url": "",
      "post_date": "04/01/2020 08:26:11",
      "content": "<p>@Moshel, by \"from the training set\", do you mean the 500G-training set? or the small sample training set \"input\\deepfake-detection-challenge\\train_sample_videos\" ? Thanks</p>",
      "votes": null,
      "replies": [
        {
          "id": 793823,
          "author_name": "moshel",
          "author_url": "",
          "post_date": "04/01/2020 10:14:05",
          "content": "<p>From the 500gb. You can find the Metadata files is some datasets. Look at the public kernels. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "793612": "For further research of deepfake detection, we make request to ground truth answer of test videos from deepfake detection challenge and evaluate performance of new models in the future.",
    "793635": "Do you mean the 400 videos? They are from the training set so you have the gt.",
    "793726": "Moshel, by \"from the training set\", do you mean the 500G-training set? or the small sample training set \"input\\deepfake-detection-challenge\\train_sample_videos\" ? Thanks",
    "793823": "From the 500gb. You can find the Metadata files is some datasets. Look at the public kernels.",
    "794727": "Thank you very much! I mean the 400 videos from input/deepfake-detection-challenge/test_videos. We predict these videos and finish the submission.csv.",
    "794741": "They are videos from the big training set, from folders 0 and 1 iirc. You can get the gt from the training metadata",
    "794748": "OK. I check the big training set. Thank you!"
  },
  "source": "meta"
}