You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To check how well your method/models can generalize,
I tried using ./scripts/kitti15_save.sh to infer disparity maps on new and previously unseen data,
but unfortunately save_disp.py is too dependant on the KITTI data structure (e.g. via --testlist)
and I also do not know what the "third column" in ./filenames/*.txt is used for,
otherwise, I could maybe hack a script myself.
E.g. from ./filenames/kitti15_train.txt: training/image_2/000000_10.png training/image_3/000000_10.png training/disp_occ_0/000000_10.png
left right ???
To check how well your method/models can generalize,
I tried using
./scripts/kitti15_save.sh
to infer disparity maps on new and previously unseen data,but unfortunately
save_disp.py
is too dependant on the KITTI data structure (e.g. via--testlist
)and I also do not know what the "third column" in
./filenames/*.txt
is used for,otherwise, I could maybe hack a script myself.
E.g. from
./filenames/kitti15_train.txt
:training/image_2/000000_10.png training/image_3/000000_10.png training/disp_occ_0/000000_10.png
left right ???
Having something like this would be great:
(Assuming the corresponding file names within
$PATH_TO_LEFT_IMAGES_FOLDER
and$PATH_TO_RIGHT_IMAGES_FOLDER
are the same.)The text was updated successfully, but these errors were encountered: