{
  "id": 545009,
  "title": "General Approach",
  "url": "/competitions/bhf-data-science-centre-ecg-challenge/discussion/545009",
  "author_name": "Rodrigo M Carrillo Larco",
  "post_date": "2024-11-08T02:49:17.901000",
  "votes": 0,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello all,</p>\n<p>I wonder what general approach you are using given the large size of images and not so much memory? </p>\n<p>For example, I started testing simple models and then transfer learning with regular architectures such as VGG and DenseNet. However, I've been trying with a small subset to get quick results and iterate. Even in this case it takes a couple of hours. So I have not yet reached a very decent solution. Thanks for any advice!</p>\n<p>Best and good luck!</p>",
  "messages": [
    {
      "id": 3044946,
      "postDate": "2024-11-14T02:30:14.370Z",
      "content": "<p>If you have limited computational resources, I recommend using smaller models and subsets for quick validation. Also given the larger image size, resize can save you time, which usually doesn't significantly impact performance.</p>",
      "rawMarkdown": "If you have limited computational resources, I recommend using smaller models and subsets for quick validation. Also given the larger image size, resize can save you time, which usually doesn't significantly impact performance.",
      "votes": 1
    },
    {
      "id": 3039404,
      "postDate": "2024-11-08T02:49:17.900Z",
      "content": "<p>Hello all,</p>\n<p>I wonder what general approach you are using given the large size of images and not so much memory? </p>\n<p>For example, I started testing simple models and then transfer learning with regular architectures such as VGG and DenseNet. However, I've been trying with a small subset to get quick results and iterate. Even in this case it takes a couple of hours. So I have not yet reached a very decent solution. Thanks for any advice!</p>\n<p>Best and good luck!</p>",
      "rawMarkdown": "Hello all,\n\nI wonder what general approach you are using given the large size of images and not so much memory? \n\nFor example, I started testing simple models and then transfer learning with regular architectures such as VGG and DenseNet. However, I've been trying with a small subset to get quick results and iterate. Even in this case it takes a couple of hours. So I have not yet reached a very decent solution. Thanks for any advice!\n\nBest and good luck!"
    }
  ],
  "comments": [
    {
      "id": 3044946,
      "author_name": "Zhongli Wu",
      "author_url": "",
      "post_date": "2024-11-14T02:30:14.370000",
      "content": "<p>If you have limited computational resources, I recommend using smaller models and subsets for quick validation. Also given the larger image size, resize can save you time, which usually doesn't significantly impact performance.</p>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3044946": "If you have limited computational resources, I recommend using smaller models and subsets for quick validation. Also given the larger image size, resize can save you time, which usually doesn't significantly impact performance.",
    "3039404": "Hello all,\n\nI wonder what general approach you are using given the large size of images and not so much memory? \n\nFor example, I started testing simple models and then transfer learning with regular architectures such as VGG and DenseNet. However, I've been trying with a small subset to get quick results and iterate. Even in this case it takes a couple of hours. So I have not yet reached a very decent solution. Thanks for any advice!\n\nBest and good luck!"
  }
}