{
  "id": 174939,
  "title": "What's your best single PyTorch model?",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/174939",
  "author_name": "",
  "post_date": "2020-08-16T10:26:36.293704400Z",
  "votes": -1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Since Tensorflow seems to have an edge with TPU, I'm curious to see what people achieve with PyTorch and GPU (TPU?).</p>\n<p>Post your CV and LB score below ;).</p>\n<p>I'll update mine ASAP.</p>",
  "messages": [
    {
      "id": "972198",
      "postDate": "08/16/2020 10:26:36",
      "content": "<p>Since Tensorflow seems to have an edge with TPU, I'm curious to see what people achieve with PyTorch and GPU (TPU?).</p>\n<p>Post your CV and LB score below ;).</p>\n<p>I'll update mine ASAP.</p>",
      "rawMarkdown": "Since Tensorflow seems to have an edge with TPU, I'm curious to see what people achieve with PyTorch and GPU (TPU?).\n\nPost your CV and LB score below ;).\n\nI'll update mine ASAP.",
      "votes": null
    },
    {
      "id": "972933",
      "postDate": "08/17/2020 01:31:35",
      "content": "<p>-8.1719 LB<br>\n<a href=\"https://www.kaggle.com/jonykarki/pytorch-qr\" target=\"_blank\">https://www.kaggle.com/jonykarki/pytorch-qr</a><br>\nWhat kind of model are you using?</p>",
      "rawMarkdown": "8.1719 LB\n[https://www.kaggle.com/jonykarki/pytorch-qr](https://www.kaggle.com/jonykarki/pytorch-qr)\nWhat kind of model are you using?",
      "votes": null
    },
    {
      "id": "973218",
      "postDate": "08/17/2020 07:14:34",
      "content": "<p><a href=\"https://www.kaggle.com/jonykarki\" target=\"_blank\">@jonykarki</a> I'm reaching -6.85 (lr = 0.1, 150 epochs) with the same model, could you please share your hyperparameters ?</p>\n<p>Question to people who scored -6.7x , are you using only tabular data ?  I'm pretty scared about overfitting , public LB is not reliable since it contains at least 15% of the test set</p>",
      "rawMarkdown": "jonykarki I'm reaching -6.85 (lr = 0.1, 150 epochs) with the same model, could you please share your hyperparameters ?\n\nQuestion to people who scored -6.7x , are you using only tabular data ?  I'm pretty scared about overfitting , public LB is not reliable since it contains at least 15% of the test set",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 972933,
      "author_name": "jonykarki",
      "author_url": "",
      "post_date": "08/17/2020 01:31:35",
      "content": "<p>-8.1719 LB<br>\n<a href=\"https://www.kaggle.com/jonykarki/pytorch-qr\" target=\"_blank\">https://www.kaggle.com/jonykarki/pytorch-qr</a><br>\nWhat kind of model are you using?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 973218,
      "author_name": "alexj21",
      "author_url": "",
      "post_date": "08/17/2020 07:14:34",
      "content": "<p><a href=\"https://www.kaggle.com/jonykarki\" target=\"_blank\">@jonykarki</a> I'm reaching -6.85 (lr = 0.1, 150 epochs) with the same model, could you please share your hyperparameters ?</p>\n<p>Question to people who scored -6.7x , are you using only tabular data ?  I'm pretty scared about overfitting , public LB is not reliable since it contains at least 15% of the test set</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "972198": "Since Tensorflow seems to have an edge with TPU, I'm curious to see what people achieve with PyTorch and GPU (TPU?).\n\nPost your CV and LB score below ;).\n\nI'll update mine ASAP.",
    "972933": "8.1719 LB\n[https://www.kaggle.com/jonykarki/pytorch-qr](https://www.kaggle.com/jonykarki/pytorch-qr)\nWhat kind of model are you using?",
    "973218": "jonykarki I'm reaching -6.85 (lr = 0.1, 150 epochs) with the same model, could you please share your hyperparameters ?\n\nQuestion to people who scored -6.7x , are you using only tabular data ?  I'm pretty scared about overfitting , public LB is not reliable since it contains at least 15% of the test set"
  },
  "source": "meta"
}