{
  "id": 99763,
  "title": "Some Query regarding using Old Competition data",
  "url": "/competitions/aptos2019-blindness-detection/discussion/99763",
  "author_name": "",
  "post_date": "2019-07-14T03:35:06.513223100Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I am thinking of using old competition data ,but i was wondering how much time it would take to commit 🤔 ,given the  fact the it is taking around 3 hrs to commit the kernel using the present competition data which is around 15-20 x smaller than the old competition data.How much time is it taking for your model to commit if you are using old competition data?</p>",
  "messages": [
    {
      "id": "574532",
      "postDate": "07/14/2019 03:35:06",
      "content": "<p>I am thinking of using old competition data ,but i was wondering how much time it would take to commit 🤔 ,given the  fact the it is taking around 3 hrs to commit the kernel using the present competition data which is around 15-20 x smaller than the old competition data.How much time is it taking for your model to commit if you are using old competition data?</p>",
      "rawMarkdown": "I am thinking of using old competition data ,but i was wondering how much time it would take to commit 🤔 ,given the  fact the it is taking around 3 hrs to commit the kernel using the present competition data which is around 15-20 x smaller than the old competition data.How much time is it taking for your model to commit if you are using old competition data?",
      "votes": null
    },
    {
      "id": "574719",
      "postDate": "07/14/2019 11:36:03",
      "content": "<p>Using last comp. data, I can run around 10-1x epochs for 9 hours, with keras fit_generator ... (depending on image size ans preprocessing scheme)</p>",
      "rawMarkdown": "Using last comp. data, I can run around 10-1x epochs for 9 hours, with keras fit_generator ... (depending on image size ans preprocessing scheme)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 574719,
      "author_name": "ratthachat",
      "author_url": "",
      "post_date": "07/14/2019 11:36:03",
      "content": "<p>Using last comp. data, I can run around 10-1x epochs for 9 hours, with keras fit_generator ... (depending on image size ans preprocessing scheme)</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "574532": "I am thinking of using old competition data ,but i was wondering how much time it would take to commit 🤔 ,given the  fact the it is taking around 3 hrs to commit the kernel using the present competition data which is around 15-20 x smaller than the old competition data.How much time is it taking for your model to commit if you are using old competition data?",
    "574719": "Using last comp. data, I can run around 10-1x epochs for 9 hours, with keras fit_generator ... (depending on image size ans preprocessing scheme)"
  },
  "source": "meta"
}