How to implement FP16?
Closed this issue · 2 comments
408550969 commented
Thank you for sharing the code. I have a few questions:
- How to implement FP16?
- Why wait for several seconds when running the inference and display the following lines:
---- Might be an error below, Error ----
Came till here
Whats happening here ?
Is it here 1
Is it here 2
Is it here 2.1
Half2 support requested on hardware without native FP16 support, performance will be negatively affected.
@Saikumargadde @JingliangGao
JingliangGao commented
Maybe it depends on your hardware . An example can be given is that my nvidia TX2 support FP16 , but server can not .
408550969 commented
Thanks