{
  "id": 328686,
  "title": "Amex Metric in Pytorch",
  "url": "/competitions/amex-default-prediction/discussion/328686",
  "author_name": "",
  "post_date": "2022-06-02T13:10:32.201021600Z",
  "votes": 3,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Has anyone implemented metric in Pytorch?  Asking for a friend.</p>",
  "messages": [
    {
      "id": "1809163",
      "postDate": "06/02/2022 13:10:32",
      "content": "<p>Has anyone implemented metric in Pytorch?  Asking for a friend.</p>",
      "rawMarkdown": "Has anyone implemented metric in Pytorch?  Asking for a friend.",
      "votes": null
    },
    {
      "id": "1809176",
      "postDate": "06/02/2022 13:21:29",
      "content": "<p>I'm not sure why you need it in PyTorch, but here is a numpy implementation: <a href=\"https://www.kaggle.com/competitions/amex-default-prediction/discussion/327534\" target=\"_blank\">link</a></p>",
      "rawMarkdown": "I'm not sure why you need it in PyTorch, but here is a numpy implementation: [link](https://www.kaggle.com/competitions/amex-default-prediction/discussion/327534)",
      "votes": null
    },
    {
      "id": "1809500",
      "postDate": "06/02/2022 18:07:33",
      "content": "<p>I wanted to use it for EarlyStopping. Is there a way to use this version for it?</p>",
      "rawMarkdown": "I wanted to use it for EarlyStopping. Is there a way to use this version for it?",
      "votes": null
    },
    {
      "id": "1810035",
      "postDate": "06/03/2022 07:51:06",
      "content": "<p>I've shared a PyTorch implementation in <a href=\"https://www.kaggle.com/code/rohanrao/amex-competition-metric-implementations#PyTorch-Implementation\" target=\"_blank\">my notebook</a> along with 5 more implementations (Pandas, Datatable, JAX, Numpy, TensorFlow) including cross-verification and benchmarking along them.</p>",
      "rawMarkdown": "I've shared a PyTorch implementation in [my notebook](https://www.kaggle.com/code/rohanrao/amex-competition-metric-implementations#PyTorch-Implementation) along with 5 more implementations (Pandas, Datatable, JAX, Numpy, TensorFlow) including cross-verification and benchmarking along them.",
      "votes": null
    },
    {
      "id": "1810151",
      "postDate": "06/03/2022 10:02:06",
      "content": "<p>Thanks Rohan! That was what i needed</p>",
      "rawMarkdown": "Thanks Rohan! That was what i needed",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1809176,
      "author_name": "bacicnikola",
      "author_url": "",
      "post_date": "06/02/2022 13:21:29",
      "content": "<p>I'm not sure why you need it in PyTorch, but here is a numpy implementation: <a href=\"https://www.kaggle.com/competitions/amex-default-prediction/discussion/327534\" target=\"_blank\">link</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 1809500,
          "author_name": "huseyincot",
          "author_url": "",
          "post_date": "06/02/2022 18:07:33",
          "content": "<p>I wanted to use it for EarlyStopping. Is there a way to use this version for it?</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1810035,
      "author_name": "rohanrao",
      "author_url": "",
      "post_date": "06/03/2022 07:51:06",
      "content": "<p>I've shared a PyTorch implementation in <a href=\"https://www.kaggle.com/code/rohanrao/amex-competition-metric-implementations#PyTorch-Implementation\" target=\"_blank\">my notebook</a> along with 5 more implementations (Pandas, Datatable, JAX, Numpy, TensorFlow) including cross-verification and benchmarking along them.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1810151,
          "author_name": "huseyincot",
          "author_url": "",
          "post_date": "06/03/2022 10:02:06",
          "content": "<p>Thanks Rohan! That was what i needed</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1809163": "Has anyone implemented metric in Pytorch?  Asking for a friend.",
    "1809176": "I'm not sure why you need it in PyTorch, but here is a numpy implementation: [link](https://www.kaggle.com/competitions/amex-default-prediction/discussion/327534)",
    "1809500": "I wanted to use it for EarlyStopping. Is there a way to use this version for it?",
    "1810035": "I've shared a PyTorch implementation in [my notebook](https://www.kaggle.com/code/rohanrao/amex-competition-metric-implementations#PyTorch-Implementation) along with 5 more implementations (Pandas, Datatable, JAX, Numpy, TensorFlow) including cross-verification and benchmarking along them.",
    "1810151": "Thanks Rohan! That was what i needed"
  },
  "source": "meta"
}