Google Colab or Kaggle notebook?

This post is based on my experience of using both, I would keep on adding any new details.

Google colab vs Kaggle

I have been using Google Colab over Kaggle only because of these reasons which are very strong

1. Colab doesn't have a limit of GPU usage quota like Kaggle has of 30 hr per week but you can only use Colab GPU straight for 9 hours, after that you need to reload it
2. Big advantage of Colab is you can open more than one notebook and can use GPU on them simultaneously, while Kaggle allows only one notebook to use GPU, so training different models become easy on Colab
3. You can load a notebook straight from a GitHub account and commit back to it, all versions will be saved in the GitHub.
4. You can mount your own data which is stored in the drive to colab environment and can work upon it
5. All your Jupyter notebooks are stored in the drive

Some nuisance you might face for using Colab instead of Kaggle

1. For Kaggle competitions you don't have to download data every time, but for using Colab you need to download the data every time in the session

For the best method to download data in the Colab, would be putting the next article soon with the link here.

The above article would keep on updating depending upon my experience in both