I want to do something like optimize operator or implement new operators.
But I’m new to pytorch, I watch the code like how pytorch implement relu. and I’m confused how can pytorch call the cpp file from python file?
I 'm working at developing some operators in CUDA I want to publish a paper about my job. But I’m new about this, and my professor is new,too. So If you can give some advise I will be very very thankful!
Thanks for your reply!