{
  "id": 46881,
  "title": "Pretrained classical CV models, like Resnets are not allowed?",
  "url": "/competitions/tensorflow-speech-recognition-challenge/discussion/46881",
  "author_name": "",
  "post_date": "2018-01-04T12:49:00.295430300Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I saw that using external data, including pretrained weights data is not allowed for this competition. I understand this limitation as  preventive measure to combat, for example audio models that can distinguish different words and theoretically could help with classifying unknown unknowns etc.   </p>\n\n<p>However classical pretrained on ImageNet CV models, like Inceptions and Resnets, have no idea about these audio labels. Are these Imagenet pretrained weights are also NOT allowed ?</p>\n\n<p>I would appreciate if someone from admins can confirm this .</p>",
  "messages": [
    {
      "id": "265038",
      "postDate": "01/04/2018 12:49:00",
      "content": "<p>I saw that using external data, including pretrained weights data is not allowed for this competition. I understand this limitation as  preventive measure to combat, for example audio models that can distinguish different words and theoretically could help with classifying unknown unknowns etc.   </p>\n\n<p>However classical pretrained on ImageNet CV models, like Inceptions and Resnets, have no idea about these audio labels. Are these Imagenet pretrained weights are also NOT allowed ?</p>\n\n<p>I would appreciate if someone from admins can confirm this .</p>",
      "rawMarkdown": "I saw that using external data, including pretrained weights data is not allowed for this competition. I understand this limitation as  preventive measure to combat, for example audio models that can distinguish different words and theoretically could help with classifying unknown unknowns etc.   \n\nHowever classical pretrained on ImageNet CV models, like Inceptions and Resnets, have no idea about these audio labels. Are these Imagenet pretrained weights are also NOT allowed ?\n\nI would appreciate if someone from admins can confirm this .",
      "votes": null
    },
    {
      "id": "265114",
      "postDate": "01/04/2018 16:42:55",
      "content": "<p>Correct, they are not allowed. We are not differentiating between different types of pre-trained models.</p>",
      "rawMarkdown": "Correct, they are not allowed. We are not differentiating between different types of pre-trained models.",
      "votes": null
    },
    {
      "id": "265136",
      "postDate": "01/04/2018 17:35:23",
      "content": "<p>Thank you.</p>",
      "rawMarkdown": "Thank you.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 265114,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "01/04/2018 16:42:55",
      "content": "<p>Correct, they are not allowed. We are not differentiating between different types of pre-trained models.</p>",
      "votes": null,
      "replies": [
        {
          "id": 265136,
          "author_name": "heyt0ny",
          "author_url": "",
          "post_date": "01/04/2018 17:35:23",
          "content": "<p>Thank you.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "265038": "I saw that using external data, including pretrained weights data is not allowed for this competition. I understand this limitation as  preventive measure to combat, for example audio models that can distinguish different words and theoretically could help with classifying unknown unknowns etc.   \n\nHowever classical pretrained on ImageNet CV models, like Inceptions and Resnets, have no idea about these audio labels. Are these Imagenet pretrained weights are also NOT allowed ?\n\nI would appreciate if someone from admins can confirm this .",
    "265114": "Correct, they are not allowed. We are not differentiating between different types of pre-trained models.",
    "265136": "Thank you."
  },
  "source": "meta"
}