{
  "id": 101220,
  "title": "Note:truncnorm is giving uplift to score - good work by sakami",
  "url": "/competitions/generative-dog-images/discussion/101220",
  "author_name": "",
  "post_date": "2019-07-24T05:09:12.439811900Z",
  "votes": 6,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I tried the trick of truncnorm at the end while generating output , mentioned in the kernel below by @sakami </p>\n\n<p><a href=\"https://www.kaggle.com/sakami/ralsgan-dogs-cropping-random\">https://www.kaggle.com/sakami/ralsgan-dogs-cropping-random</a></p>\n\n<p>This gave me a 2-point update (65 to 63)  without changing anything else in the kernel . So n00bs like me have a look at the same while trying .</p>\n\n<p>I am using DCGAN (unlike the RaLSGAN mentioned in the kernel )</p>",
  "messages": [
    {
      "id": "583147",
      "postDate": "07/24/2019 05:09:12",
      "content": "<p>I tried the trick of truncnorm at the end while generating output , mentioned in the kernel below by @sakami </p>\n\n<p><a href=\"https://www.kaggle.com/sakami/ralsgan-dogs-cropping-random\">https://www.kaggle.com/sakami/ralsgan-dogs-cropping-random</a></p>\n\n<p>This gave me a 2-point update (65 to 63)  without changing anything else in the kernel . So n00bs like me have a look at the same while trying .</p>\n\n<p>I am using DCGAN (unlike the RaLSGAN mentioned in the kernel )</p>",
      "rawMarkdown": "I tried the trick of truncnorm at the end while generating output , mentioned in the kernel below by @sakami \n\nhttps://www.kaggle.com/sakami/ralsgan-dogs-cropping-random\n\nThis gave me a 2-point update (65 to 63)  without changing anything else in the kernel . So n00bs like me have a look at the same while trying .\n\nI am using DCGAN (unlike the RaLSGAN mentioned in the kernel )",
      "votes": null
    },
    {
      "id": "583245",
      "postDate": "07/24/2019 08:03:26",
      "content": "<p>It's my pleasure. ;)</p>",
      "rawMarkdown": "It's my pleasure. ;)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 583245,
      "author_name": "sakami",
      "author_url": "",
      "post_date": "07/24/2019 08:03:26",
      "content": "<p>It's my pleasure. ;)</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "583147": "I tried the trick of truncnorm at the end while generating output , mentioned in the kernel below by @sakami \n\nhttps://www.kaggle.com/sakami/ralsgan-dogs-cropping-random\n\nThis gave me a 2-point update (65 to 63)  without changing anything else in the kernel . So n00bs like me have a look at the same while trying .\n\nI am using DCGAN (unlike the RaLSGAN mentioned in the kernel )",
    "583245": "It's my pleasure. ;)"
  },
  "source": "meta"
}