The miou of the pretrained model on semantickitty is only about 58.97.
Xavier-wa opened this issue · comments
Nice Work! But I have some problem about the miou of the pretrained model on semantickitty,when I upload the result of evalution,It is only 58.98 but not 59.5
my results on semantickitti seq 08 is 0.558?
my results on semantickitti seq 08 is 0.558?
you can modify the KNN["use"] to True and try again
my results on semantickitti seq 08 is Acc avg 0.016 IoU avg 0.005 with parameters and ./eval.sh -d /mnt/500_SSD/dataset/semantic-kitti/ -m pretrained -p preds -s valid -c 30
postproc parameters
post:
CRF:
use: False
train: True
params: False # this should be a dict when in use
KNN:
use: False # This parameter default is false
params:
knn: 5
search: 5
sigma: 1.0
cutoff: 1.0