{
  "id": 16817,
  "title": "Advice for newbies",
  "url": "/competitions/diabetic-retinopathy-detection/discussion/16817",
  "author_name": "BhavyaaSharma",
  "post_date": "2015-10-04T18:03:22.870000",
  "votes": 3,
  "comment_count": 0,
  "views": 873,
  "content": "<p>I developed an interest in this competition and I am trying to attempt the same on python. I was finally able to download the data but I do not have access to a GPU. I have the following questions for now:-\n1. how much of my performance is going to be affected by not attempting it on a GPU\n2. The train dataset has around 35k images or so, how should I go about the pre-processing the entire dataset swiftly? Are there any example codes that could help me with this? Some of the NN codes have a bit on transforming the dataset, is there a way to carry out transformation and augmentation together?</p>",
  "messages": [
    {
      "id": 95055,
      "postDate": "2015-10-04T18:03:22.870Z",
      "content": "<p>I developed an interest in this competition and I am trying to attempt the same on python. I was finally able to download the data but I do not have access to a GPU. I have the following questions for now:-\n1. how much of my performance is going to be affected by not attempting it on a GPU\n2. The train dataset has around 35k images or so, how should I go about the pre-processing the entire dataset swiftly? Are there any example codes that could help me with this? Some of the NN codes have a bit on transforming the dataset, is there a way to carry out transformation and augmentation together?</p>",
      "rawMarkdown": "I developed an interest in this competition and I am trying to attempt the same on python. I was finally able to download the data but I do not have access to a GPU. I have the following questions for now:-\r\n1. how much of my performance is going to be affected by not attempting it on a GPU\r\n2. The train dataset has around 35k images or so, how should I go about the pre-processing the entire dataset swiftly? Are there any example codes that could help me with this? Some of the NN codes have a bit on transforming the dataset, is there a way to carry out transformation and augmentation together?",
      "votes": 3
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "95055": "I developed an interest in this competition and I am trying to attempt the same on python. I was finally able to download the data but I do not have access to a GPU. I have the following questions for now:-\r\n1. how much of my performance is going to be affected by not attempting it on a GPU\r\n2. The train dataset has around 35k images or so, how should I go about the pre-processing the entire dataset swiftly? Are there any example codes that could help me with this? Some of the NN codes have a bit on transforming the dataset, is there a way to carry out transformation and augmentation together?"
  }
}