Thomas Tse
Results
1
comments of
Thomas Tse
I noticed significant slow down using cuda (quadro p1000) too. Using model_size=medium.en, compute_type="in8" Downgrading to 1.0.2 resolved this issue.