Conversation
Signed-off-by: Wenqi Li <[email protected]>
Signed-off-by: Wenqi Li <[email protected]>
|
pending the PR, I will test with spleen fast training and get back ASAP. Thanks. |
|
Hi @rijobro , BTW, have you implemented this feature locally in your branch? If yes, maybe you can share some comments here. Thanks. |
Signed-off-by: Wenqi Li <[email protected]>
0428d66 to
9415e43
Compare
|
Mark: training time with this PR around Thanks. |
I looked into this issue, the root cause is that MONAI/monai/transforms/utils.py Line 427 in 28fb7d6 and (2) the numbers are converted into tensors: MONAI/monai/transforms/croppad/array.py Line 405 in 28fb7d6 Previously So I think the issue is not really in this PR, it's from the |
|
Hi @wyli , OK, thanks for your investigation! Thanks. |
thanks! please help fix the issue and we can merge this PR after that |
|
Sure, sounds good to me. Investigating. Thanks. |
|
This looks great, thanks! |
|
I think we already found the root cause of the performance issue, no need to put the Thanks. |
Signed-off-by: Wenqi Li [email protected]
Fixes #2975
Description
unravel_index is still in numpy, otherwise it supports both backendsadded both backendsStatus
Ready
Types of changes
./runtests.sh -f -u --net --coverage../runtests.sh --quick --unittests.make htmlcommand in thedocs/folder.