{
  "id": 131688,
  "title": "Pls dont post high scoring kernel ,just strategy still fine",
  "url": "/competitions/deepfake-detection-challenge/discussion/131688",
  "author_name": "",
  "post_date": "2020-02-21T03:00:12.143065300Z",
  "votes": 4,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I request to those who are going to publish high scoring kernel not to post submission  kernel that can score more than 420,410 at this stage  . It is absolutely fine to give a strategy to go past these scores but posting kernel with submission and weights is an easy escalator  to those who dont put niche efforts in the competitions . this also prevents them  from learning or applying new things. </p>",
  "messages": [
    {
      "id": "752444",
      "postDate": "02/21/2020 03:00:12",
      "content": "<p>I request to those who are going to publish high scoring kernel not to post submission  kernel that can score more than 420,410 at this stage  . It is absolutely fine to give a strategy to go past these scores but posting kernel with submission and weights is an easy escalator  to those who dont put niche efforts in the competitions . this also prevents them  from learning or applying new things. </p>",
      "rawMarkdown": "I request to those who are going to publish high scoring kernel not to post submission  kernel that can score more than 420,410 at this stage  . It is absolutely fine to give a strategy to go past these scores but posting kernel with submission and weights is an easy escalator  to those who dont put niche efforts in the competitions . this also prevents them  from learning or applying new things.",
      "votes": null
    },
    {
      "id": "752525",
      "postDate": "02/21/2020 05:45:33",
      "content": "<p>+1👍 I was wanting to post this topic too! People need to figure they're own way out for that top 10%.</p>",
      "rawMarkdown": "1👍 I was wanting to post this topic too! People need to figure they're own way out for that top 10%.",
      "votes": null
    },
    {
      "id": "752554",
      "postDate": "02/21/2020 06:37:01",
      "content": "<p>In my opinion, the inference kernel is very helpful. But learned weights do not helpful at all.</p>\n\n<p>However if they don't write down their high LB scores, nobody click notebook and give 👍. :P</p>",
      "rawMarkdown": "In my opinion, the inference kernel is very helpful. But learned weights do not helpful at all.\n\nHowever if they don't write down their high LB scores, nobody click notebook and give 👍. :P",
      "votes": null
    },
    {
      "id": "753005",
      "postDate": "02/21/2020 15:58:54",
      "content": "<p>inference kernels are ENOUGH. There's A TON of inference kernels and it's pointless to post MORE since they all share the same concept.</p>",
      "rawMarkdown": "inference kernels are ENOUGH. There's A TON of inference kernels and it's pointless to post MORE since they all share the same concept.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 752525,
      "author_name": "unkownhihi",
      "author_url": "",
      "post_date": "02/21/2020 05:45:33",
      "content": "<p>+1👍 I was wanting to post this topic too! People need to figure they're own way out for that top 10%.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 752554,
      "author_name": "caffeinism",
      "author_url": "",
      "post_date": "02/21/2020 06:37:01",
      "content": "<p>In my opinion, the inference kernel is very helpful. But learned weights do not helpful at all.</p>\n\n<p>However if they don't write down their high LB scores, nobody click notebook and give 👍. :P</p>",
      "votes": null,
      "replies": [
        {
          "id": 753005,
          "author_name": "unkownhihi",
          "author_url": "",
          "post_date": "02/21/2020 15:58:54",
          "content": "<p>inference kernels are ENOUGH. There's A TON of inference kernels and it's pointless to post MORE since they all share the same concept.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "752444": "I request to those who are going to publish high scoring kernel not to post submission  kernel that can score more than 420,410 at this stage  . It is absolutely fine to give a strategy to go past these scores but posting kernel with submission and weights is an easy escalator  to those who dont put niche efforts in the competitions . this also prevents them  from learning or applying new things.",
    "752525": "1👍 I was wanting to post this topic too! People need to figure they're own way out for that top 10%.",
    "752554": "In my opinion, the inference kernel is very helpful. But learned weights do not helpful at all.\n\nHowever if they don't write down their high LB scores, nobody click notebook and give 👍. :P",
    "753005": "inference kernels are ENOUGH. There's A TON of inference kernels and it's pointless to post MORE since they all share the same concept."
  },
  "source": "meta"
}