This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

[Numpy] Implement numpy operator 'average' #16720

Merged

haojin2 merged 5 commits into apache:master from anexpn:np_op_average

Dec 9, 2019

Contributor

anexpn commented Nov 4, 2019 •

edited

Loading

Description

Implement the numpy operator 'average'

https://numpy.org/devdocs/reference/generated/numpy.average.html

Checklist

Essentials

Please feel free to remove inapplicable items for your PR.

Changes are complete (i.e. I finished coding on this PR)
All changes have test coverage:
Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
Code is well-documented:
For user-facing API changes, API doc string has been updated.
For new C++ functions in header files, their functionalities and arguments are documented.
For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
Check the API doc at https://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change

Changes

op average, tests, (and when applicable, API doc)

Comments

does not guarantee the same behavior with numpy when given float16 dtype and overflow occurs
a and weights must be the same dtype

anexpn requested review from anirudh2290, eric-haibin-lin and szha as code owners

November 4, 2019 08:57

Contributor Author

anexpn commented Nov 4, 2019

haojin2 added the Numpy label

haojin2 reviewed

View reviewed changes

include/mxnet/tensor_blob.h Outdated Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

python/mxnet/ndarray/numpy/_op.py Outdated Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

src/operator/numpy/np_broadcast_reduce_op.h Outdated Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

src/operator/numpy/np_broadcast_reduce_op.h Outdated Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

src/operator/numpy/np_broadcast_reduce_op_value.cc Outdated Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

tests/python/unittest/test_numpy_op.py Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

tests/python/unittest/test_numpy_op.py Outdated Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

src/operator/numpy/np_broadcast_reduce_op.h Outdated Show resolved Hide resolved

haojin2 reviewed

View reviewed changes

src/operator/numpy/np_broadcast_reduce_op.h Outdated Show resolved Hide resolved

anexpn commented

View reviewed changes

python/mxnet/symbol/numpy/_symbol.py Show resolved Hide resolved

anexpn commented

View reviewed changes

python/mxnet/symbol/numpy/_symbol.py Show resolved Hide resolved

anexpn commented

View reviewed changes

python/mxnet/numpy/multiarray.py Show resolved Hide resolved

python/mxnet/numpy/multiarray.py Show resolved Hide resolved

python/mxnet/ndarray/numpy/_op.py Show resolved Hide resolved

python/mxnet/ndarray/numpy/_op.py Show resolved Hide resolved

anexpn force-pushed the np_op_average branch from 0ab6fa6 to 24cec76 Compare

November 4, 2019 11:00

haojin2 approved these changes

View reviewed changes

Contributor

haojin2 left a comment

LGTM @reminisce Any extra comments?

reminisce reviewed

View reviewed changes

include/mxnet/tensor_blob.h Outdated Show resolved Hide resolved

include/mxnet/tensor_blob.h Outdated Show resolved Hide resolved

tests/python/unittest/test_numpy_op.py Outdated Show resolved Hide resolved

reminisce force-pushed the np_op_average branch from 6836a67 to e23d8c9 Compare

November 20, 2019 05:48

anexpn force-pushed the np_op_average branch 5 times, most recently from 1ae8dcc to 4ed05fa Compare

November 26, 2019 17:43


          Complete forward and backward

396f0af

Add tests

Implement returned

Implement none weights returned

Replace reduce implementation with workspace version

Temp space can be requested only once.

Use FNumVisibleOutputs attribute, always return sum of weights

Fix bug

Fix GPU bug

Fix GPU bug

Formatting

Apply suggestions from code review

two blank lines between python functions

Fix pylint

Fix GPU bug

Add support for null and addTo reqs

Remove TBlob constructors

Refactor

anexpn force-pushed the np_op_average branch from 55fd10c to 396f0af Compare

November 27, 2019 08:48

anexpn added 4 commits

November 28, 2019 21:22


          Retrigger

da2fe8d


          Retrigger

a08b704


          Retrigger

edc6484


          Retrigger

Contributor Author

anexpn commented Dec 8, 2019

@haojin2 ready to be merged

haojin2 merged commit 44cd63e into apache:master

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

Reviewers

reminisce reminisce left review comments

haojin2 haojin2 approved these changes

anirudh2290 Awaiting requested review from anirudh2290

eric-haibin-lin Awaiting requested review from eric-haibin-lin

szha Awaiting requested review from szha szha is a code owner

Labels