swc-17
Results
3
comments of
swc-17
> > Probably because the rotate-center used in bevformer is not at (0, 0). > > Any way, when adding tempral fusion(length-que = 3), I have a runtime error, have...
I found that reducing batch size per GPU from 10 to 8 will accelerate training to ~7 hours.
It seems that annotations for v1.0-test is also not complete.