-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add MedNext implementation #8004
Conversation
Signed-off-by: Suraj Pai <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the work! I think it's a great addition to MONAI!
Can't it be usefull to adapt the factory functions of create_mednext_v1 script inside your |
…edNext variants (S, B, M, L) + integration of remarks from @johnzilke (Project-MONAI#8004 (review)) for renaming class arguments - removal of self defined LayerNorm - linked residual connection for encoder and decoder Signed-off-by: Robin CREMESE <[email protected]>
Thanks for the comments @johnzielke and thank you for the updates @rcremese. I'll update my branch to dev and look through the changes in your PR asap. |
…edNext variants (S, B, M, L) + integration of remarks from @johnzilke (Project-MONAI#8004 (review)) for renaming class arguments - removal of self defined LayerNorm - linked residual connection for encoder and decoder Signed-off-by: Robin CREMESE <[email protected]>
…edNext variants (S, B, M, L) + integration of remarks from @johnzilke (Project-MONAI#8004 (review)) for renaming class arguments - removal of self defined LayerNorm - linked residual connection for encoder and decoder Signed-off-by: Robin CREMESE <[email protected]>
Signed-off-by: Suraj Pai <[email protected]>
for more information, see https://pre-commit.ci
Signed-off-by: Suraj Pai <[email protected]>
@KumoLiu This implementation should be ready now. Please let me know if you have any comments |
Signed-off-by: Suraj Pai <[email protected]>
@KumoLiu Added. Do you think there would be interest to add this as a candidate for Auto3DSeg? I refer to this paper for its performance benchmarking: https://arxiv.org/abs/2404.09556 |
Thank you for bringing this up, it's an interesting suggestion. I believe it could be worthwhile to consider this as a potential candidate for Auto3DSeg. However, before moving forward, I would appreciate hearing others' thoughts and insights on whether this aligns with the current goals and roadmap for Auto3DSeg. cc @mingxin-zheng @dongyang0122 @Nic-Ma @myron |
Great work @surajpaib ! any plans on when we're gonna be able to get this into main? |
Hi @KumoLiu, any update on merging this? We can maybe have a separate issue to discuss Auto3DSeg integration then |
/build |
Hi @surajpaib, I noticed there are still several open comments. Do you plan to address these, or would it be easier just to respond if you don't intend to make updates? Thanks! |
/build |
/build |
Fixes #7786
Description
Added MedNext architectures implementation for MONAI.
Since a lot of the code is heavily sourced from the original MedNext repo, https://github.com/MIC-DKFZ/MedNeXt, I wanted to check if there is an attribution policy with regarded to borrowed source code. I've added a derivative notice bellow the monai copyright comment. Let me know if this needs to be changed.
The blocks have been taken almost as is but the network implementation has been changed largely to allow flexible blocks and follow MONAI segresnet styling.
Types of changes
./runtests.sh -f -u --net --coverage
../runtests.sh --quick --unittests --disttests
.make html
command in thedocs/
folder.