{
  "id": 56525,
  "title": "Pretrained word embeddings as text features",
  "url": "/competitions/avito-demand-prediction/discussion/56525",
  "author_name": "",
  "post_date": "2018-05-10T16:21:37.630934Z",
  "votes": 14,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I added <a href=\"https://www.kaggle.com/christofhenkel/fasttest-common-crawl-russian\">pretrained fasttext embeddings</a> to the kaggle datasets and created a <a href=\"https://www.kaggle.com/christofhenkel/fasttext-starter-description-only\">starter kernel</a> which shows exemplary how to use them on the item description.</p>\n\n<p>Cheers, and happy kaggling.</p>",
  "messages": [
    {
      "id": "326997",
      "postDate": "05/10/2018 16:21:37",
      "content": "<p>I added <a href=\"https://www.kaggle.com/christofhenkel/fasttest-common-crawl-russian\">pretrained fasttext embeddings</a> to the kaggle datasets and created a <a href=\"https://www.kaggle.com/christofhenkel/fasttext-starter-description-only\">starter kernel</a> which shows exemplary how to use them on the item description.</p>\n\n<p>Cheers, and happy kaggling.</p>",
      "rawMarkdown": "I added [pretrained fasttext embeddings][1] to the kaggle datasets and created a [starter kernel][2] which shows exemplary how to use them on the item description.\n\nCheers, and happy kaggling.\n\n\n  [1]: https://www.kaggle.com/christofhenkel/fasttest-common-crawl-russian\n  [2]: https://www.kaggle.com/christofhenkel/fasttext-starter-description-only",
      "votes": null
    },
    {
      "id": "327849",
      "postDate": "05/12/2018 16:49:06",
      "content": "<p>Thanks, that helps!</p>",
      "rawMarkdown": "Thanks, that helps!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 327849,
      "author_name": "tunguz",
      "author_url": "",
      "post_date": "05/12/2018 16:49:06",
      "content": "<p>Thanks, that helps!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "326997": "I added [pretrained fasttext embeddings][1] to the kaggle datasets and created a [starter kernel][2] which shows exemplary how to use them on the item description.\n\nCheers, and happy kaggling.\n\n\n  [1]: https://www.kaggle.com/christofhenkel/fasttest-common-crawl-russian\n  [2]: https://www.kaggle.com/christofhenkel/fasttext-starter-description-only",
    "327849": "Thanks, that helps!"
  },
  "source": "meta"
}