Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
Help
#641
by
Aarifkhan
- opened
My models Precision is F32 what should I select
Hi!
I'd suggest bfloat16, as it's got the same range as float32 (but a lower precision), and should convert well. (See more info on the wikipedia page).
clefourrier
changed discussion status to
closed