Precision BASE in ModelOptimizer

Precision BASE in ModelOptimizer

Hi,

In model optimizer in the supported precisions, there are BASE, FP16, FP32 etc.

For FP32 I am using the data type float and for FP16 I am using half.

Can you explain the precision BASE and what data type I need to use in programs for the same ?.

 

Thanks.

2 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Hi,

This is an internal type and shouldn't be exposed to customers- thanks for noticing and raising out attention. 

an issue was opened to remove this in future releases. 

Please use FP32 instead... 

Regards,

Stav

Leave a Comment

Please sign in to add a comment. Not a member? Join today