{
  "id": 100915,
  "title": "Online Training Platform for Models",
  "url": "/competitions/recursion-cellular-image-classification/discussion/100915",
  "author_name": "Fairoza",
  "post_date": "2019-07-22T04:30:09.051000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>What type of online training platform do use to train the models?</p>\n\n<p>I didn’t get to request for bonus GCP credits, so I’m considering of using Colab. Is colab powerful enough for this data? Or maybe I should consider sign up for GCP and use it. Does GCP works the same with colab or more similar to AWS? \nI’ve tried to train using Kaggle kernel and the lowest batch size I can use is 16, which is not very good I think.</p>",
  "messages": [
    {
      "id": 581543,
      "postDate": "2019-07-22T04:30:09.050Z",
      "content": "<p>What type of online training platform do use to train the models?</p>\n\n<p>I didn’t get to request for bonus GCP credits, so I’m considering of using Colab. Is colab powerful enough for this data? Or maybe I should consider sign up for GCP and use it. Does GCP works the same with colab or more similar to AWS? \nI’ve tried to train using Kaggle kernel and the lowest batch size I can use is 16, which is not very good I think.</p>",
      "rawMarkdown": "What type of online training platform do use to train the models?\n\nI didn’t get to request for bonus GCP credits, so I’m considering of using Colab. Is colab powerful enough for this data? Or maybe I should consider sign up for GCP and use it. Does GCP works the same with colab or more similar to AWS? \nI’ve tried to train using Kaggle kernel and the lowest batch size I can use is 16, which is not very good I think.",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "581543": "What type of online training platform do use to train the models?\n\nI didn’t get to request for bonus GCP credits, so I’m considering of using Colab. Is colab powerful enough for this data? Or maybe I should consider sign up for GCP and use it. Does GCP works the same with colab or more similar to AWS? \nI’ve tried to train using Kaggle kernel and the lowest batch size I can use is 16, which is not very good I think."
  }
}