Skip to content
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.

Sign up for GitHub

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

Jump to bottom

merge master #3

Merged
merged 19 commits into from
Jan 11, 2021
Merged

merge master #3

merged 19 commits into from
Jan 11, 2021

Conversation

Nic-Ma
Copy link
Owner

@Nic-Ma Nic-Ma commented Jan 11, 2021

Fixes # .

Description

A few sentences describing the changes proposed in this pull request.

Status

Ready/Work in progress/Hold

Types of changes

  • Non-breaking change (fix or new feature that would not break existing functionality).
  • Breaking change (fix or new feature that would cause existing functionality to change).
  • New tests added to cover the changes.
  • Integration tests passed locally by running ./runtests.sh --codeformat --coverage.
  • Quick tests passed locally by running ./runtests.sh --quick.
  • In-line docstrings updated.
  • Documentation updated, tested make html command in the docs/ folder.

Nic-Ma and others added 19 commits December 17, 2020 11:59
@Nic-Ma
[DLMED] add missing import (#1376)
09638a7 
Signed-off-by: Nic Ma <[email protected]>
@wyli
adds jpeg (#1379)
1608b10 
Signed-off-by: Wenqi Li <[email protected]>
@Nic-Ma
[DLMED] add missing arg in GANTrainer (#1383)
b7ecd7b 
Signed-off-by: Nic Ma <[email protected]>
@charliebudd
1198 efficient differentiable bilateral filter (#1375)
70be9ab 
* Initial cpu implementation

Signed-off-by: charliebudd <[email protected]>

* hue speed increase moving away from torch tensors

Signed-off-by: charliebudd <[email protected]>

* allowing generic channel dimension size

Signed-off-by: charliebudd <[email protected]>

* allowing generic dimensionality

Signed-off-by: charliebudd <[email protected]>

* manual testing script for bilateral filter

Signed-off-by: charliebudd <[email protected]>

* Initial commit of CRF using permutohedral lattice to optimise message passing, implementation from https://github.com/SamuelJoutard/Permutohedral_attention_module

Signed-off-by: charliebudd <[email protected]>

* temporary example scripts demonstrating the usage of crf and phl, and some data to process

Signed-off-by: charliebudd <[email protected]>

* alternate phl implementation using a more efficient cuda hashtable, code currently untested due to windows build issues, implementation from https://github.com/SamuelJoutard/Permutohedral_attention_module

Signed-off-by: charliebudd <[email protected]>

* c++ cpu permutohedral lattice implementation

Signed-off-by: charliebudd <[email protected]>

* comparison script

Signed-off-by: charliebudd <[email protected]>

* cleaning

Signed-off-by: charliebudd <[email protected]>

* backwards pass

Signed-off-by: charliebudd <[email protected]>

* initial cuda brute force kernel

Signed-off-by: charliebudd <[email protected]>

* reverting setup.py

Signed-off-by: charliebudd <[email protected]>

* working cuda kernel for brute force 2d kernel

Signed-off-by: charliebudd <[email protected]>

* fixing unresolved symbols when compiling without cuda

Signed-off-by: charliebudd <[email protected]>

* removing macros file

Signed-off-by: charliebudd <[email protected]>

* placeholder cuda phl implementation

Signed-off-by: charliebudd <[email protected]>

* changing test case color sigma

Signed-off-by: charliebudd <[email protected]>

* Initial import of reference cuda phl implementation

Signed-off-by: charliebudd <[email protected]>

* fixing cuda kernel color weight error

Signed-off-by: charliebudd <[email protected]>

* initial edits to cuda phl code

Signed-off-by: charliebudd <[email protected]>

* fixing errors in cuda phl implementation

Signed-off-by: charliebudd <[email protected]>

* removing logging from cuda phl

Signed-off-by: charliebudd <[email protected]>

* using template data_ptr function

Signed-off-by: charliebudd <[email protected]>

* fixed cpu phl implementation

Signed-off-by: charliebudd <[email protected]>

* updating cpu phl to run on arbitrary input dimensions and channles

Signed-off-by: charliebudd <[email protected]>

* 2d and 3d testing scripts

Signed-off-by: charliebudd <[email protected]>

* Generalising permutohedral implementation to remove independance on width and height

Signed-off-by: charliebudd <[email protected]>

* removing width and height from permutohedral filter function and implementing templated feature creation

Signed-off-by: charliebudd <[email protected]>

* fixed typo

Signed-off-by: charliebudd <[email protected]>

* exteneding 3d testing script

Signed-off-by: charliebudd <[email protected]>

* generalising bruteforce cuda implementation

Signed-off-by: charliebudd <[email protected]>

* updating testing scripts

Signed-off-by: charliebudd <[email protected]>

* file organisiation

Signed-off-by: charliebudd <[email protected]>

* fixing weight error in cpu bruteforce implementation, also inlineing functions

Signed-off-by: charliebudd <[email protected]>

* some refactoring and introducing proper batch handling

Signed-off-by: charliebudd <[email protected]>

* fixing indexing error at border

Signed-off-by: charliebudd <[email protected]>

* fixing some artifacts in cuda phl for high color sigmas

Signed-off-by: charliebudd <[email protected]>

* fixing gaussian kernel function for cpu and cuda bruteforce

Signed-off-by: charliebudd <[email protected]>

* ensuring kernel is an odd numbered size

Signed-off-by: charliebudd <[email protected]>

* adding tests for precised implementation

Signed-off-by: charliebudd <[email protected]>

* adding approximate imlpementation test

Signed-off-by: charliebudd <[email protected]>

* templating implementations based on scalar_type

Signed-off-by: charliebudd <[email protected]>

* cleaning up

Signed-off-by: charliebudd <[email protected]>

* code formatting

Signed-off-by: charliebudd <[email protected]>

* removing half precision implementation for permutohedral due to errors caused on windows

Signed-off-by: charliebudd <[email protected]>

* skipping cuda tests if cuda missing

Signed-off-by: charliebudd <[email protected]>

* reformating

Signed-off-by: charliebudd <[email protected]>

* adding unit test skip based on cpp extention availablility

Signed-off-by: charliebudd <[email protected]>

* removing unused import

Signed-off-by: charliebudd <[email protected]>

* changing to  use of THCatomic add

Signed-off-by: charliebudd <[email protected]>

* adding missing licenses

Signed-off-by: charliebudd <[email protected]>

* clang reformat

Signed-off-by: charliebudd <[email protected]>
@rijobro
update contributing and normalis* -> normaliz* (#1388)
36b8302 
Signed-off-by: Richard Brown <[email protected]>
@Nic-Ma
[DLMED] add output_dtype (#1394)
f20e132 
Signed-off-by: Nic Ma <[email protected]>
@rijobro
with eval_mode(net): (#1384)
a4b1144 
* eval_mode

Signed-off-by: Richard Brown <[email protected]>

* DynUnetSupervision unittest shouldnt use eval_mode

Signed-off-by: Richard Brown <[email protected]>

* move to device

Signed-off-by: Richard Brown <[email protected]>
@Nic-Ma
1390 Enhance DynUNet with get_feature_maps API (#1393)
e2c0ed2 
* [DLMED] add get_feature_maps API to dynunet

Signed-off-by: Nic Ma <[email protected]>

* [DLMED] update according to comments

Signed-off-by: Nic Ma <[email protected]>
@wyli
adds missing desc (#1403)
0c2411e 
Signed-off-by: Wenqi Li <[email protected]>
@wyli @Nic-Ma @IsaacYangSLA
part of #926 (#1368)
1924b43 
Signed-off-by: Wenqi Li <[email protected]>

Co-authored-by: Nic Ma <[email protected]>
Co-authored-by: Isaac Yang <[email protected]>
@yiheng-wang-nv
Implement dice cross entropy loss (#1409)
115f146 
Signed-off-by: yiheng-wang-nv <[email protected]>
@rijobro
Unify visualisation (#1381)
a74fef9 
* unify visualisation
@kate-sann5100
1405 add bending energy loss (#1408)
5aa5e16 
* 1405 add bending energy loss

Signed-off-by: kate-sann5100 <[email protected]>
@danielschulz
Fix failures from many pulls from hub.docker.com (Done) (#1406)
db831ae 
* changed year 2020 in copyright to "2020 - 2021" for both occurences

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* added SHA256 check before unpacking downloaded artifacts
added list of Nvidia maintainers
resolves: "docker image tests failed due to dockerhub limitation #1380"

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* added danielschulz-specific changes for GitHu Actions to work properly

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* added test

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* revoked test line

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* migrated code to local registry instead of hub.docker.com to circumvent rate limiting on it
remark: do not use it as pull-through cache as this would not work with static labels and docker images would diverge from one another

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* reverted back local-only changes

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* improved checking integrity of NGC CLI

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* removed typo in GH repository variable

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* added remark

Signed-off-by: Daniel Schulz [email protected] <[email protected]>

* improved MAINTAINER label
got rid of "as base" extension in FROM -- seems not be be pickup downwards (is no multi-stage Docker build -- hence: removed)

Signed-off-by: Daniel Schulz [email protected] <[email protected]>
@wyli
fixes dockerfile (#1413)
db7bf89 
Signed-off-by: Wenqi Li <[email protected]>
@rijobro
Wildcard import (#1411)
1f74ca2 
add `__all__` and `remove import *`. Update CONTRIBUTING.md
@wyli @IsaacYangSLA
Fixes ci (#1416)
53c145d 
* temp tests

Signed-off-by: Wenqi Li <[email protected]>

* fixes typo

Signed-off-by: Wenqi Li <[email protected]>

* temp tests

Signed-off-by: Wenqi Li <[email protected]>

* fixes docker login warning

Signed-off-by: Wenqi Li <[email protected]>

* Revert "temp tests"

This reverts commit 99c6d20.

Signed-off-by: Wenqi Li <[email protected]>

* Revert "temp tests"

This reverts commit c529bf9.

Signed-off-by: Wenqi Li <[email protected]>

Co-authored-by: Isaac Yang <[email protected]>
@IsaacYangSLA
Improve docker build process (#1419)
a4ef691 
* Improve docker build process, fixes #1276
@kate-sann5100
1412 local normalized cross correlation (#1415)
1544d0c 
* 1412 add local normalized cross correlation

Signed-off-by: kate-sann5100 <[email protected]>
@Nic-Ma Nic-Ma merged commit 60639e3 into Nic-Ma:1420-enhance-infer-doc Jan 11, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
8 participants
@Nic-Ma @wyli @yiheng-wang-nv @charliebudd @rijobro @kate-sann5100 @danielschulz @IsaacYangSLA