Skip to content
This repository was archived by the owner on Nov 17, 2023. It is now read-only.

Add alias for np.__version__ and np.dtype #17777

Merged
reminisce merged 1 commit into from
Mar 6, 2020
Merged

Add alias for np.__version__ and np.dtype #17777

reminisce merged 1 commit into from
Mar 6, 2020

Conversation

@haojin2
Copy link
Contributor

@haojin2 haojin2 commented Mar 5, 2020

Description

As title.

Checklist

Essentials

Please feel free to remove inapplicable items for your PR.

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

Changes

  • Feature1, tests, (and when applicable, API doc)
  • Feature2, tests, (and when applicable, API doc)

Comments

  • If this change is a backward incompatible change, why must this change be made.
  • Interesting edge cases to note here

@haojin2 haojin2 added the Numpy label Mar 5, 2020
@haojin2 haojin2 requested a review from reminisce March 5, 2020 19:07
@haojin2 haojin2 requested a review from szha as a code owner March 5, 2020 19:07
@haojin2 haojin2 self-assigned this Mar 5, 2020
leezu
leezu reviewed Mar 5, 2020
View reviewed changes
python/mxnet/numpy/fallback.py
'vander',
]

__version__ = onp.__version__
Copy link
Contributor

@leezu leezu Mar 5, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is there any mechanism to ensure mx.np is compatible with the onp.__version__? If not, should this return the version mx.np is compatible with?

Copy link
Contributor Author

@haojin2 haojin2 Mar 5, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

At least for now the CI have 1.18 installed and we pass all current tests against that version. So, I think we're in good shape here.

Copy link
Contributor

@leezu leezu Mar 5, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That's great. My question is, once Numpy 1.19 or later versions are released and potentially change API, should mx.np.__version__ report 1.18, which is the version you are currently targeting, or should it claim 1.19 which could be the numpy version installed on the users machine.

Copy link
Contributor Author

@haojin2 haojin2 Mar 5, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Once a 1.19 comes out our CI would directly report problems if any and we can fix it, I cannot predict what they're changing in the future, and the dilemma here is that the fallback-ed operators will follow the new 1.19.

Copy link
Contributor

@leezu leezu Mar 5, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for clarifying the issue. How should this be handled for stable releases? Users may install the release at any time and the numpy version could be incompatible? Should the supported numpy version be explicitly called out and returned by mx.np.__version__?

Copy link
Contributor Author

@haojin2 haojin2 Mar 5, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it should be addressed by setting the correct dependencies in our release version. If we claim that we are compatible with 1.xx version but the pypi packages says that anything up to 1.yy is okay, then I think it's something that whoever is in charge of CD should address.

@reminisce reminisce merged commit cfb474b into apache:master Mar 6, 2020
MoisesHer pushed a commit to MoisesHer/incubator-mxnet that referenced this pull request Apr 10, 2020
@haojin2 @MoisesHer
add alias for np.__version__, np._NoValue and np.dtype (apache#17777)
7e2d696
anirudh2290 pushed a commit to anirudh2290/mxnet that referenced this pull request May 29, 2020
@haojin2 @anirudh2290
add alias for np.__version__, np._NoValue and np.dtype (apache#17777)
5c6f175
sxjscience pushed a commit to sxjscience/mxnet that referenced this pull request Jul 1, 2020
@haojin2 @sxjscience
add alias for np.__version__, np._NoValue and np.dtype (apache#17777)
f00c3b5
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Reviewers

@reminisce reminisce Awaiting requested review from reminisce

@szha szha Awaiting requested review from szha szha is a code owner

1 more reviewer

@leezu leezu leezu left review comments

Reviewers whose approvals may not affect merge requirements

Assignees

@haojin2 haojin2

Labels

Numpy

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@haojin2 @leezu @reminisce