about eval and test
Closed this issue · 3 comments
xuchen-dev commented
xuchen-dev commented
klingner commented
I am not quite sure what you are aiming at. Results on different datasets are of course different and not comparable. A better result on one dataset mainly tells you that the dataset is easier for the model. In the paper mainly results on kitti_zhou_test are reported as is common.
xuchen-dev commented
thank you