{
  "id": 228503,
  "title": " Huggingface Question Answering",
  "url": "/competitions/coleridgeinitiative-show-us-the-data/discussion/228503",
  "author_name": "",
  "post_date": "2021-03-25T00:06:35.093521500Z",
  "votes": 6,
  "comment_count": 1,
  "views": 0,
  "content": "<p>This is not exactly what the competition metric is asking for, but is an interesting experiment nonetheless.</p>\n<p>I've taken the Huggingface Question Answering pre-trained model, and asked it to predict which dataset is referenced (as opposed to the text mentioning it).</p>\n<p>Results can be seen in the notebook. </p>\n<ul>\n<li><a href=\"https://www.kaggle.com/jamesmcguigan/coleridge-huggingface-question-answering/\" target=\"_blank\">https://www.kaggle.com/jamesmcguigan/coleridge-huggingface-question-answering/</a></li>\n</ul>\n<p>I still have some technical issues with figuring out how to get to run in offline mode without an internet connection (can't submit to the competition without this).</p>\n<p>I am also seeking advice on how best to fine-tune a Huggingface model against a dataset.</p>",
  "messages": [
    {
      "id": "1251604",
      "postDate": "03/25/2021 00:06:35",
      "content": "<p>This is not exactly what the competition metric is asking for, but is an interesting experiment nonetheless.</p>\n<p>I've taken the Huggingface Question Answering pre-trained model, and asked it to predict which dataset is referenced (as opposed to the text mentioning it).</p>\n<p>Results can be seen in the notebook. </p>\n<ul>\n<li><a href=\"https://www.kaggle.com/jamesmcguigan/coleridge-huggingface-question-answering/\" target=\"_blank\">https://www.kaggle.com/jamesmcguigan/coleridge-huggingface-question-answering/</a></li>\n</ul>\n<p>I still have some technical issues with figuring out how to get to run in offline mode without an internet connection (can't submit to the competition without this).</p>\n<p>I am also seeking advice on how best to fine-tune a Huggingface model against a dataset.</p>",
      "rawMarkdown": "This is not exactly what the competition metric is asking for, but is an interesting experiment nonetheless.\n\nI've taken the Huggingface Question Answering pre-trained model, and asked it to predict which dataset is referenced (as opposed to the text mentioning it).\n\nResults can be seen in the notebook. \n- https://www.kaggle.com/jamesmcguigan/coleridge-huggingface-question-answering/\n\nI still have some technical issues with figuring out how to get to run in offline mode without an internet connection (can't submit to the competition without this).\n\nI am also seeking advice on how best to fine-tune a Huggingface model against a dataset.",
      "votes": null
    },
    {
      "id": "1252909",
      "postDate": "03/26/2021 07:07:39",
      "content": "<p>this approach was the very first approach I wanted to test. Thank, you did it.</p>",
      "rawMarkdown": "this approach was the very first approach I wanted to test. Thank, you did it.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1252909,
      "author_name": "masoudmzb",
      "author_url": "",
      "post_date": "03/26/2021 07:07:39",
      "content": "<p>this approach was the very first approach I wanted to test. Thank, you did it.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1251604": "This is not exactly what the competition metric is asking for, but is an interesting experiment nonetheless.\n\nI've taken the Huggingface Question Answering pre-trained model, and asked it to predict which dataset is referenced (as opposed to the text mentioning it).\n\nResults can be seen in the notebook. \n- https://www.kaggle.com/jamesmcguigan/coleridge-huggingface-question-answering/\n\nI still have some technical issues with figuring out how to get to run in offline mode without an internet connection (can't submit to the competition without this).\n\nI am also seeking advice on how best to fine-tune a Huggingface model against a dataset.",
    "1252909": "this approach was the very first approach I wanted to test. Thank, you did it."
  },
  "source": "meta"
}