{
  "id": 90495,
  "title": "For fast.ai users, kernel with simpler model ready",
  "url": "/competitions/freesound-audio-tagging-2019/discussion/90495",
  "author_name": "",
  "post_date": "2019-04-24T09:29:13.196339600Z",
  "votes": 7,
  "comment_count": 5,
  "views": 0,
  "content": "<p><a href=\"https://www.kaggle.com/daisukelab/cnn-2d-basic-3-using-simple-model\">https://www.kaggle.com/daisukelab/cnn-2d-basic-3-using-simple-model</a></p>\n\n<p>This uses simple model borrowing from the kernel below:\n<a href=\"https://www.kaggle.com/mhiro2/simple-2d-cnn-classifier-with-pytorch\">https://www.kaggle.com/mhiro2/simple-2d-cnn-classifier-with-pytorch</a>\n(link was wrong, fixed)</p>\n\n<ul>\n<li>fast.ai ready, you can easily customize.</li>\n<li>A little bit improved training parameters, which still leaves space to improve ;)</li>\n</ul>",
  "messages": [
    {
      "id": "522350",
      "postDate": "04/24/2019 09:29:13",
      "content": "<p><a href=\"https://www.kaggle.com/daisukelab/cnn-2d-basic-3-using-simple-model\">https://www.kaggle.com/daisukelab/cnn-2d-basic-3-using-simple-model</a></p>\n\n<p>This uses simple model borrowing from the kernel below:\n<a href=\"https://www.kaggle.com/mhiro2/simple-2d-cnn-classifier-with-pytorch\">https://www.kaggle.com/mhiro2/simple-2d-cnn-classifier-with-pytorch</a>\n(link was wrong, fixed)</p>\n\n<ul>\n<li>fast.ai ready, you can easily customize.</li>\n<li>A little bit improved training parameters, which still leaves space to improve ;)</li>\n</ul>",
      "rawMarkdown": "https://www.kaggle.com/daisukelab/cnn-2d-basic-3-using-simple-model\n\nThis uses simple model borrowing from the kernel below:\nhttps://www.kaggle.com/mhiro2/simple-2d-cnn-classifier-with-pytorch\n(link was wrong, fixed)\n\n- fast.ai ready, you can easily customize.\n- A little bit improved training parameters, which still leaves space to improve ;)",
      "votes": null
    },
    {
      "id": "522392",
      "postDate": "04/24/2019 11:29:45",
      "content": "<p><strong>Thanks for sharing <a href=\"/daisukelab\">@daisukelab</a></strong> </p>",
      "rawMarkdown": "**Thanks for sharing @daisukelab**",
      "votes": null
    },
    {
      "id": "522496",
      "postDate": "04/24/2019 13:56:39",
      "content": "<p>Thanks for sharing!</p>",
      "rawMarkdown": "Thanks for sharing!",
      "votes": null
    },
    {
      "id": "522731",
      "postDate": "04/24/2019 22:53:29",
      "content": "<p>Thanks for sharing! Out of curiosity, do you know why the public score is lower for fastai vs the pytorch version?</p>",
      "rawMarkdown": "Thanks for sharing! Out of curiosity, do you know why the public score is lower for fastai vs the pytorch version?",
      "votes": null
    },
    {
      "id": "522796",
      "postDate": "04/25/2019 03:29:05",
      "content": "<p>Thanks for your simple model. :)</p>",
      "rawMarkdown": "Thanks for your simple model. :)",
      "votes": null
    },
    {
      "id": "522797",
      "postDate": "04/25/2019 03:31:27",
      "content": "<p>Tuning hp and some other detail would make that difference.\nYou can either choose basic pytorch or fast.ai implementation to start improvement for you.</p>",
      "rawMarkdown": "Tuning hp and some other detail would make that difference.\nYou can either choose basic pytorch or fast.ai implementation to start improvement for you.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 522392,
      "author_name": "akashravichandran",
      "author_url": "",
      "post_date": "04/24/2019 11:29:45",
      "content": "<p><strong>Thanks for sharing <a href=\"/daisukelab\">@daisukelab</a></strong> </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 522496,
      "author_name": "mhiro2",
      "author_url": "",
      "post_date": "04/24/2019 13:56:39",
      "content": "<p>Thanks for sharing!</p>",
      "votes": null,
      "replies": [
        {
          "id": 522796,
          "author_name": "daisukelab",
          "author_url": "",
          "post_date": "04/25/2019 03:29:05",
          "content": "<p>Thanks for your simple model. :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 522731,
      "author_name": "tanlikesmath",
      "author_url": "",
      "post_date": "04/24/2019 22:53:29",
      "content": "<p>Thanks for sharing! Out of curiosity, do you know why the public score is lower for fastai vs the pytorch version?</p>",
      "votes": null,
      "replies": [
        {
          "id": 522797,
          "author_name": "daisukelab",
          "author_url": "",
          "post_date": "04/25/2019 03:31:27",
          "content": "<p>Tuning hp and some other detail would make that difference.\nYou can either choose basic pytorch or fast.ai implementation to start improvement for you.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "522350": "https://www.kaggle.com/daisukelab/cnn-2d-basic-3-using-simple-model\n\nThis uses simple model borrowing from the kernel below:\nhttps://www.kaggle.com/mhiro2/simple-2d-cnn-classifier-with-pytorch\n(link was wrong, fixed)\n\n- fast.ai ready, you can easily customize.\n- A little bit improved training parameters, which still leaves space to improve ;)",
    "522392": "**Thanks for sharing @daisukelab**",
    "522496": "Thanks for sharing!",
    "522731": "Thanks for sharing! Out of curiosity, do you know why the public score is lower for fastai vs the pytorch version?",
    "522796": "Thanks for your simple model. :)",
    "522797": "Tuning hp and some other detail would make that difference.\nYou can either choose basic pytorch or fast.ai implementation to start improvement for you."
  },
  "source": "meta"
}