pytorch precision float Lightning supports either double precision (64), ... I make the boxes to be flaot32 but when I output them using dataloaders, ...
確定! 回上一頁