{
  "id": 458663,
  "title": "27th Tabtransformer",
  "url": "/competitions/open-problems-single-cell-perturbations/writeups/iraqbot-27th-tabtransformer",
  "author_name": "",
  "post_date": "2023-12-01T10:22:42.727Z",
  "votes": 7,
  "comment_count": 2,
  "views": 0,
  "content": "<p>🫡 to the shake up survivor</p>\n<p>Used ensemble of ft transformer of 12 models(same as tab because no numerical) trained on different seed, 5 folds by models all used</p>\n<p>config picked based on mae (dim range 48 to 64, depth 4 to 8, head 8)<br>\nVery high dropout 0.4<br>\nNo augmentation, no use of the blend ressources (i just checked one and i was lost)</p>\n<p>Hard time to understand how to start and the competition<br>\nI did not take the competition seriously until i dropped from the top 10 of the public leaderboard because of the blend fog, but was too lazy to do/think anything fancy to get back on track.</p>\n<p>1st tabular competition (i am not a fan of that!)</p>\n<p>bronze to silver next goal is gold !</p>",
  "messages": [
    {
      "id": "2544613",
      "postDate": "12/01/2023 01:14:53",
      "content": "<p>🫡 to the shake up survivor</p>\n<p>Used ensemble of ft transformer of 12 models(same as tab because no numerical) trained on different seed, 5 folds by models all used</p>\n<p>config picked based on mae (dim range 48 to 64, depth 4 to 8, head 8)<br>\nVery high dropout 0.4<br>\nNo augmentation, no use of the blend ressources (i just checked one and i was lost)</p>\n<p>Hard time to understand how to start and the competition<br>\nI did not take the competition seriously until i dropped from the top 10 of the public leaderboard because of the blend fog, but was too lazy to do/think anything fancy to get back on track.</p>\n<p>1st tabular competition (i am not a fan of that!)</p>\n<p>bronze to silver next goal is gold !</p>",
      "rawMarkdown": "🫡 to the shake up survivor\n\nUsed ensemble of ft transformer of 12 models(same as tab because no numerical) trained on different seed, 5 folds by models all used\n\nconfig picked based on mae (dim range 48 to 64, depth 4 to 8, head 8)\nVery high dropout 0.4\nNo augmentation, no use of the blend ressources (i just checked one and i was lost)\n\nHard time to understand how to start and the competition\nI did not take the competition seriously until i dropped from the top 10 of the public leaderboard because of the blend fog, but was too lazy to do/think anything fancy to get back on track.\n\n1st tabular competition (i am not a fan of that!)\n\nbronze to silver next goal is gold !",
      "votes": null
    },
    {
      "id": "2544741",
      "postDate": "12/01/2023 04:03:22",
      "content": "<p>Congratulations. Thanks for sharing the details. </p>",
      "rawMarkdown": "Congratulations. Thanks for sharing the details.",
      "votes": null
    },
    {
      "id": "2696657",
      "postDate": "03/14/2024 13:37:57",
      "content": "<p>Congratulations with medal on Kaggle Single Cell Perturbations Challenge ! <br>\nI wonder would it be possible to share the code of the solutions ? </p>",
      "rawMarkdown": "Congratulations with medal on Kaggle Single Cell Perturbations Challenge ! \nI wonder would it be possible to share the code of the solutions ?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2544741,
      "author_name": "crsuthikshnkumar",
      "author_url": "",
      "post_date": "12/01/2023 04:03:22",
      "content": "<p>Congratulations. Thanks for sharing the details. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2696657,
      "author_name": "alexandervc",
      "author_url": "",
      "post_date": "03/14/2024 13:37:57",
      "content": "<p>Congratulations with medal on Kaggle Single Cell Perturbations Challenge ! <br>\nI wonder would it be possible to share the code of the solutions ? </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2544613": "🫡 to the shake up survivor\n\nUsed ensemble of ft transformer of 12 models(same as tab because no numerical) trained on different seed, 5 folds by models all used\n\nconfig picked based on mae (dim range 48 to 64, depth 4 to 8, head 8)\nVery high dropout 0.4\nNo augmentation, no use of the blend ressources (i just checked one and i was lost)\n\nHard time to understand how to start and the competition\nI did not take the competition seriously until i dropped from the top 10 of the public leaderboard because of the blend fog, but was too lazy to do/think anything fancy to get back on track.\n\n1st tabular competition (i am not a fan of that!)\n\nbronze to silver next goal is gold !",
    "2544741": "Congratulations. Thanks for sharing the details.",
    "2696657": "Congratulations with medal on Kaggle Single Cell Perturbations Challenge ! \nI wonder would it be possible to share the code of the solutions ?"
  },
  "source": "meta"
}