{
  "id": 117361,
  "title": "Best Kaggle kernel solutions",
  "url": "/competitions/rsna-intracranial-hemorrhage-detection/discussion/117361",
  "author_name": "Michal",
  "post_date": "2019-11-14T23:02:12.623000",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Is there anyone in stage 2 who used only Kaggle kernels? I wonder what was possible to achieve without having your own hardware or using services like GCP. With the GPU limitations image classification/segmentation competitions are getting pretty challenging for people without the resources so hopefully we can see that it's not all that bad and you can still get good results.</p>\n\n<p>Please share your best score in stage 2 you've been able to get using Kaggle kernels and possibly some info about your solution.</p>\n\n<p>Congratulations to all winners and many thanks to Kaggle, RSNA, ASNR and MD.ai for hosting such a great and interesting competition.</p>",
  "messages": [
    {
      "id": 673386,
      "postDate": "2019-11-14T23:02:12.623Z",
      "content": "<p>Is there anyone in stage 2 who used only Kaggle kernels? I wonder what was possible to achieve without having your own hardware or using services like GCP. With the GPU limitations image classification/segmentation competitions are getting pretty challenging for people without the resources so hopefully we can see that it's not all that bad and you can still get good results.</p>\n\n<p>Please share your best score in stage 2 you've been able to get using Kaggle kernels and possibly some info about your solution.</p>\n\n<p>Congratulations to all winners and many thanks to Kaggle, RSNA, ASNR and MD.ai for hosting such a great and interesting competition.</p>",
      "rawMarkdown": "Is there anyone in stage 2 who used only Kaggle kernels? I wonder what was possible to achieve without having your own hardware or using services like GCP. With the GPU limitations image classification/segmentation competitions are getting pretty challenging for people without the resources so hopefully we can see that it's not all that bad and you can still get good results.\n\nPlease share your best score in stage 2 you've been able to get using Kaggle kernels and possibly some info about your solution.\n\nCongratulations to all winners and many thanks to Kaggle, RSNA, ASNR and MD.ai for hosting such a great and interesting competition.",
      "votes": 2
    },
    {
      "id": 675607,
      "postDate": "2019-11-18T10:08:38.990Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 675607,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-11-18T10:08:38.990000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "673386": "Is there anyone in stage 2 who used only Kaggle kernels? I wonder what was possible to achieve without having your own hardware or using services like GCP. With the GPU limitations image classification/segmentation competitions are getting pretty challenging for people without the resources so hopefully we can see that it's not all that bad and you can still get good results.\n\nPlease share your best score in stage 2 you've been able to get using Kaggle kernels and possibly some info about your solution.\n\nCongratulations to all winners and many thanks to Kaggle, RSNA, ASNR and MD.ai for hosting such a great and interesting competition.",
    "675607": ""
  }
}