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

Fix Deformable Convolutional Op Support #179 #1056

Closed
wants to merge 614 commits into from
Closed

Fix Deformable Convolutional Op Support #179 #1056

wants to merge 614 commits into from

Conversation

smallsunsun1
Copy link

Add cpu and gpu implementation for Deformable Conv2d and DeformablePSROIAlign according to https://arxiv.org/abs/1811.11168

shun-lin and others added 30 commits October 30, 2019 15:30
@shun-lin @seanpmorgan
Adds TQDMProgressBar in callbacks (#610)
3783529 
* added TQDMProgressBar and tutorial
@shun-lin @seanpmorgan
[Small typo] fixed typo in tqdm_progress_bar colab tutorial (#653)
29af552
@mels630 @WindQAQ
Add unit test verifying tf.Dataset support for dense_image_warp (#332) (
34cb568 
#654)
@shun-lin @Squadrick
update _toc.yaml with path to TQDM Progress Bar colab (#657)
c2c7020
@rahulunair @seanpmorgan
Update Readme, add self as poincare maintainer (#667)
00446ee
@rahulunair @seanpmorgan
[doc]: Add self to skip_gram_ops maintainer list (#666)
d6648d3
@rahulunair @seanpmorgan
Update README.md - add self to maintainer (#665)
db21011
@Squadrick @seanpmorgan
CLN: Refactor f_scores and f_test (#502)
0cc04e2 
* CLN: Refactor f_scores and f_test

* Add `threshold` param to f-scores
* Tests now compare with sklearn
* Add sklearn to requirements

* Format files

* Add F1 score test

* Register FBetaScore and F1Score as Keras custom objects
* Update readme to separate both metrics

* Add test for F1-score get_config

* FIX: Use sk_score for true value

* Remove sklearn from f_test

Resort to using hard coded test cases
rather than comparing with sklearn

* Remove unused import

* Rename test_keras_model -> _get_model
@Squadrick @seanpmorgan
FIX: Internally cast to required DType (#659)
cd1d4e7 
* FIX: Internally cast to required DType
* Add pylint disable for unscriptable-object bug
@seanpmorgan @WindQAQ
[WIP] Use public Keras object registration (#669)
b7be70a 
* Remove test skips since upstream fix

* Use public keras object registration

* Merge master and update registration

* Fix tf import

* Update READMEs

* F scores keras registration

* Lint
@WindQAQ @seanpmorgan
Fix overflow of CohenKappa (#675)
87b2db3 
* fix overflow of int32
@shun-lin @seanpmorgan
add get_config to tqdm (#681)
1151d75
@autoih @WindQAQ
Fixing missing files (#678)
a778ebc 
* fix links

* missing import lamb

* reorder
@Squadrick @WindQAQ
Set random seed for TensorFlow to make test repro (#685)
3d18335
@Squadrick @seanpmorgan
FIX: Invariant input_spec for WeightNormalization (#687)
45a6f5f 
* Make the first dimension `None` to support invariant batch size.
* Add test case to check compatibility of WeightNormalization with
  TimeDistributed.
@seanpmorgan
Add cyclical learning rate schedulers (#644)
f421dd2 
* Add cyclical learning rate schedulers
@Squadrick @seanpmorgan
Build data_init layer under name_scope (#691)
5fd7501 
* Build data_init layer under name_scope

The original wrapped layer and the non-trainable layer created for data
dependent initialization had a clash in their namespaces. Creating the
second layer under a name scope of 'data_dep_init' fixes the issue.

* Lint

* Add test for saving

* Use create_tempfile
@tzachar @seanpmorgan
Fix wrapped layer has use_biase=False (#696)
fea7878
@WindQAQ @seanpmorgan
Speed up correlation cost (#689)
2060ff0 
* sharding over pixel

* robust test on cpu

* fix typo
@0xshreyash @Squadrick
Add Stochastic Weight Averaging Optimizer. (#620)
baffaa9
@autoih @WindQAQ
MatthewsCorrelationCoefficient (#483)
98ebef7 
* add mcc py

* update & test file

* test file revision

* indention

* revise

* build file

* change dtype

* remove type

* correct numerator multiplication

* code format check

* format

* minor

* minor

* modify doc

* sample weight

* import

* layers

* avoid using get_shape

* multi-class for true_negative

* correct true negative

* updae README and add test case

* minor fixing multi-lines

* output dtype

* move docstring to exact place and keep data type as optional

* minor change using tf api

* revision

* tf api and minor revision
@WindQAQ @Squadrick
Fix keras model compile for lifted_struct_loss (#700)
e553c43 
* fix keras model compile

* checkout pylint change

* disable pylint

* make linter happy
@autoih @seanpmorgan
remove copyright notice (#702)
3b8eb44
@seanpmorgan @WindQAQ
Revert pinned nightly (#658)
5972d3c
@gharibian @seanpmorgan
Migrate from std::string to tensorflow::tstring. (#705)
670699e 
Note that during the transition period tstring is typedef'ed to
std::string.

See: tensorflow/community#91
@Squadrick @seanpmorgan
Don't build parse_time till TF r2.1 (#710)
3dacb5a 
* Don't build parse_time till TF r2.1

* Fix TODO, and BUILD

* Remove .py file from build target
@fsx950223 @WindQAQ
Add GIOU loss (#477)
e1fa98b 
* Add GIOU loss

* Refact giou calculate

* Fix doc

* Update Readme.md

* Format code

* refact calculate

* fix document

* fix readme

* fix docs

* Change to official api

* format code

* enhance robust

* add box format

* add keras test

* add one bbox test

* add different shapes test case

* format code

* fix docs

* make private

* add interger test

* format code

* change expression
@WindQAQ @seanpmorgan
update documentation (#714)
7fef5fb
@autoih @WindQAQ
add resampler kernel (#662)
4d0e595 
* add resampler kernel

* add register op

* namespace and register

* python format

* headers and cleanup

* sanity cleanup

* readme update

* alphabetic order

* gpu test & minor revision

* comment on wrapping part

* cpu test

* miscellaneous fixing

* minior fix

* line removal
@seanpmorgan @WindQAQ
Hotfix: fix brew setup (#716)
e77195b
@review-notebook-app
Copy link

Check out this pull request on  ReviewNB

You'll be able to see Jupyter notebook diff and discuss changes. Powered by ReviewNB.

@googlebot
Copy link

CLAs look good, thanks!

ℹ️ Googlers: Go here for more info.

@googlebot googlebot added cla: yes and removed cla: no labels Feb 9, 2020
@seanpmorgan
Copy link
Member

Thank you very much for the contribution @smallsunsun1 ! Just wanted to comment and say we'll get to reviewing this as soon as time allows.

@googlebot
Copy link

We found a Contributor License Agreement for you (the sender of this pull request), but were unable to find agreements for all the commit author(s) or Co-authors. If you authored these, maybe you used a different email address in the git commits than was used to sign the CLA (login here to double check)? If these were authored by someone else, then they will need to sign a CLA as well, and confirm that they're okay with these being contributed to Google.
In order to pass this check, please resolve this problem and then comment @googlebot I fixed it.. If the bot doesn't comment, it means it doesn't think anything has changed.

ℹ️ Googlers: Go here for more info.

@smallsunsun1
Copy link
Author

@googlebot I fixed it.

@smallsunsun1
Copy link
Author

@seanpmorgan . I add a EquiConv according to https://arxiv.org/pdf/1903.08094.pdf. But when i commit,
I made a mistake in email address, it changed to sunjiahe@sss-2.local, original is 975759105@qq.com, then the cla test failed, how can i fixed in minimal effort.

@smallsunsun1 smallsunsun1 mentioned this pull request Feb 22, 2020
Closed
@smallsunsun1
Copy link
Author

close this and reopen a new pull request #1129 with support for EquiConv

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@facaiy facaiy Awaiting requested review from facaiy

@seanpmorgan seanpmorgan Awaiting requested review from seanpmorgan

Assignees
No one assigned
Labels
cla: no custom-ops documentation layers tutorials
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.