Computer Vision sample code in PyTorch docs has super low accuracy

vision
1

I’ve been studying this nice sample PyTorch Computer Vision code here:

https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html

The accuracy it gives is only 54%!? I’ve tried adding more than
2 epochs but still haven’t figured out how to get it at a decent level.

Any ideas? It probably would be more confidence building for
newcomers if their first PyTorch code gave a stellar accuracy! :wink:

Sincerely,

Chris