Google Cloud VM Machine: Cuda out Memory

I have recently set-up an account on Google Cloud VM Machine, and I am using the Nvidia Tesla K-80GPU. When I run my Pytorch code I am getting the following error:
enter image description here

But when I run the same code on Google Colab, it is working fine.

You might be getting a better GPU (more memory) on Colab. Does it help if you reduce your batch size?

Seems like I don’t know why the issue was coming, I reset my VM machine and re-downloaded the data and the program ran this time.