{
  "id": 6882,
  "title": "Does any one try the ImageNet model?",
  "url": "/competitions/galaxy-zoo-the-galaxy-challenge/discussion/6882",
  "author_name": "",
  "post_date": "2014-01-21T05:41:32.917Z",
  "votes": null,
  "comment_count": 2,
  "views": 2178,
  "content": "<p>I just read in <a href=\"http://www.kaggle.com/c/dogs-vs-cats/forums/t/6845/example-use-decaf-via-nolearn-for-94-accuracy\">cat &amp; dog competition</a>, they use ImageNet model directly to generate the conv feature then run random forest, which leads very impressive result.</p>\n<p>&nbsp;</p>\n<p>I am guessing the conv feature is highly abstract and context-free feature. Does any one try in this task?</p>\n<p>Since the ImageNet is similar to cat &amp; dog but not this one, I am not sure whether it works or not. I am so busy currently and don't have time to do competition, but I think the result may be interesting.</p>\n<p>&nbsp;</p>\n<p>&nbsp;</p>",
  "messages": [
    {
      "id": "37663",
      "postDate": "01/21/2014 05:41:32",
      "content": "<p>I just read in <a href=\"http://www.kaggle.com/c/dogs-vs-cats/forums/t/6845/example-use-decaf-via-nolearn-for-94-accuracy\">cat &amp; dog competition</a>, they use ImageNet model directly to generate the conv feature then run random forest, which leads very impressive result.</p>\n<p>&nbsp;</p>\n<p>I am guessing the conv feature is highly abstract and context-free feature. Does any one try in this task?</p>\n<p>Since the ImageNet is similar to cat &amp; dog but not this one, I am not sure whether it works or not. I am so busy currently and don't have time to do competition, but I think the result may be interesting.</p>\n<p>&nbsp;</p>\n<p>&nbsp;</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "38034",
      "postDate": "01/27/2014 22:21:32",
      "content": "<p>I tried that. Got ~0.135 in validation with a random forest. Galaxy images are of different character than ImageNet images, so the features don't seem to generalize that much.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "38057",
      "postDate": "01/28/2014 07:42:23",
      "content": "<p>I would be very surprised if a network trained to distinguish 1000 real world classes works on galaxies. I don't think a pretrained network will be useful in this competition.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 38034,
      "author_name": "zygmunt",
      "author_url": "",
      "post_date": "01/27/2014 22:21:32",
      "content": "<p>I tried that. Got ~0.135 in validation with a random forest. Galaxy images are of different character than ImageNet images, so the features don't seem to generalize that much.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 38057,
      "author_name": "tdomhan",
      "author_url": "",
      "post_date": "01/28/2014 07:42:23",
      "content": "<p>I would be very surprised if a network trained to distinguish 1000 real world classes works on galaxies. I don't think a pretrained network will be useful in this competition.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "37663": "",
    "38034": "",
    "38057": ""
  },
  "source": "meta"
}