Apache MXNet 1.5.1 is a patch release that includes fixes for critical bugs and performance regressions introduced with 1.5.0 release. 
As MXNet follows semver, this release includes backwards-compatible fixes only. 
New features can be introduced in the next minor release. MXNet 1.6.0 road map can be found at https://github.com/apache/incubator-mxnet/issues/15589.

Release Manager: Tao Lv (TaoLv@GitHub), Shufan Wu (juliusshufan@GitHub)
Release shepherd:

Scope

Scope is still open - please add!

GitHub issue: [Discussion] 1.5.1 Patch Release (https://github.com/apache/incubator-mxnet/issues/15613)

Issue

Proposer

PRs

Contributor(s)

Notes

WarpCTC build failure in 1.5.0

(https://github.com/apache/incubator-mxnet/issues/15612)

@samskalickyhttps://github.com/apache/incubator-mxnet/pull/15601@samskalickyDone. The fix is picked to v1.5.x branch.

Gluon Inference failed

(https://github.com/apache/incubator-mxnet/issues/15281)

@lanking520https://github.com/apache/incubator-mxnet/pull/15298@ZhennanQinAlready in 1.5.0

[MXNet 1.5.0.rc2] Issues with asnumpy() method

(https://github.com/apache/incubator-mxnet/issues/15431)

@roywei

Closed. False positive issue.

License issues need to be fixed before 1.6 release

(https://github.com/apache/incubator-mxnet/issues/15542)

@roywei

https://github.com/apache/incubator-mxnet/pull/15806

https://github.com/apache/incubator-mxnet/pull/15828

@TaoLv

@gigasquid

@roywei

We fixed the item 1, 2, 6 on v1.5.x. Item 3,4,5 are postponed to v1.6.0 release.

https://github.com/apache/incubator-mxnet/pull/15860

https://github.com/apache/incubator-mxnet/pull/16026

https://github.com/apache/incubator-mxnet/commit/52ce718a076ca56385360a9b45d6afb44b5f1e27

update julia install doc@iblis17https://github.com/apache/incubator-mxnet/pull/15609@Ksoy

Fixed on master. No need to be ported to v1.5.x.

https://mail-archives.apache.org/mod_mbox/mxnet-dev/201908.mbox/%3CCAH1G5Zpz3sAxfM4Z8GYh7BGmVHkxYy%3DPsVOxE4DE9pQdSX7rYw%40mail.gmail.com%3E

julia: bump Windows prebult binary version to v1.5.0@iblis17https://github.com/apache/incubator-mxnet/pull/15608@iblis17Fixed on master
Night test failure on CI@lebeg

https://github.com/apache/incubator-mxnet/issues/15374

http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/NightlyTestsForBinaries/detail/master/395/pipeline/

Not an issue on v1.5.x branch. Moved to 1.6.0

https://github.com/apache/incubator-mxnet/issues/15589#issuecomment-526396099

Broken Sidebar on website API for master and 1.5.0

(https://github.com/apache/incubator-mxnet/issues/15200)

@ThomasDelteil, @sandeep-krishnamurthy 

Postponed to the next release.

https://mail-archives.apache.org/mod_mbox/mxnet-dev/201908.mbox/%3C62C61072-FDC2-4F4C-8989-FDB92BA03D03%40amazon.com%3E

1.5.0 MKLDNN error "Unknown MKLDNN format"

(https://github.com/apache/incubator-mxnet/issues/15737)

@samskalickyhttps://github.com/apache/incubator-mxnet/pull/15692@xinyu-intel

Fixed on master. Note: master uses v0.20 while v1.5.x uses v0.19 of MKL-DNN.

The fix is picked to v1.5.x branch

the bug of `MXEnginePushAsyncND` and `MXEnginePushSyncND`

(https://github.com/apache/incubator-mxnet/issues/15774)

@wkcnhttps://github.com/apache/incubator-mxnet/pull/15751@wkcnDone. The fix is picked to v1.5.x.

Java examples broken with mxnet mkldnn build

(https://github.com/apache/incubator-mxnet/issues/15267)

@TaoLv, @arcadiaphyhttps://github.com/apache/incubator-mxnet/pull/15351@wuxun-zhang

Fixed on master.

https://github.com/apache/incubator-mxnet/pull/15802

Bug or highly unintuitive and undocumented behaviour corrupting NDArray data(possible race condition)[MKLDNN][C++]

(https://github.com/apache/incubator-mxnet/issues/15555)

@TaoLv, @kostayScrhttps://github.com/apache/incubator-mxnet/pull/15637@zhennanQin

Fixed on master and v1.5.x

https://github.com/apache/incubator-mxnet/pull/15803

Dropout produces wrong mask with MKL-DNN

(https://github.com/apache/incubator-mxnet/issues/15659)

@TaoLv, @matteosalhttps://github.com/apache/incubator-mxnet/pull/15697@wuxun-zhang

Fixed on master.

https://github.com/apache/incubator-mxnet/pull/15804


MKL-DNN gives wrong convolution bias gradient if weights gradient is not requested

(https://github.com/apache/incubator-mxnet/issues/15464)

@TaoLv, @matteosal

https://github.com/apache/incubator-mxnet/pull/15497

https://github.com/apache/incubator-mxnet/pull/15631

@zixuanweeei

Fixed on master

https://github.com/apache/incubator-mxnet/pull/15805

Simple_Bind failure in 1.5.0

(https://github.com/apache/incubator-mxnet/issues/15784)

@samskalickyhttps://github.com/apache/incubator-mxnet/pull/15620@haohuanw

Fixed on master

https://github.com/apache/incubator-mxnet/pull/15873

TensorRT patches@KellenSunderland

https://github.com/apache/incubator-mxnet/pull/15645

https://github.com/apache/incubator-mxnet/pull/15666

https://github.com/apache/incubator-mxnet/pull/14898

https://github.com/apache/incubator-mxnet/pull/15344

https://github.com/apache/incubator-mxnet/pull/14860


Already fixed on master. Need be picked to v1.5.x

https://github.com/apache/incubator-mxnet/pull/15874 – done

https://github.com/apache/incubator-mxnet/pull/15875 – done

https://github.com/apache/incubator-mxnet/pull/15877 – done

https://github.com/apache/incubator-mxnet/pull/16044 – done

https://github.com/apache/incubator-mxnet/pull/16043 – WIP

Benchmark doc fix

(https://github.com/apache/incubator-mxnet/pull/15769)

@ChaiBapchyahttps://github.com/apache/incubator-mxnet/pull/15769@ChaiBapchya

Fixed on both master and v1.5.x branch.

https://github.com/apache/incubator-mxnet/pull/16029

FP16 Support for C Predict API

(https://github.com/apache/incubator-mxnet/issues/14159,

https://github.com/apache/incubator-mxnet/issues/13335)

@samskalickyhttps://github.com/apache/incubator-mxnet/pull/15245@anirudh2290

PR is working in progress:

https://github.com/apache/incubator-mxnet/pull/16027

PR is closed:"This fix depends on #15118 which is a big feature, so im closing this PR for the 1.5.x branch. It will have to wait until 1.6.x".

Error when loading models with no params

(https://github.com/apache/incubator-mxnet/issues/15916)

@samskalickyhttps://github.com/apache/incubator-mxnet/pull/15917@samskalickyhttps://github.com/apache/incubator-mxnet/pull/16028

GPU OOM: topk regression in v1.5

(https://github.com/apache/incubator-mxnet/issues/15703)

@apeforest @leezuhttps://github.com/apache/incubator-mxnet/pull/15948@apeforest

The issue was fixed on the master branch. But have problem to back port the fix to v1.5.x.

https://github.com/apache/incubator-mxnet/pull/15999

Release Timeline

Following timeline is based on everything goes well.(Added some buffer time)

Step

Tasks

Goal

Actual

Comments

Code Freeze and release startTrack ongoing PRs9/3 6pm PST






Cut the release branchCheck license headers



Make code changes with necessary version updates9/3 6pm PST


Cut the release branch

from v1.5.x branch

Update the version on master


Test the release and tag the releaseNightly test, Jenkins CI9/4


RAT check

checked by CI

Tag RC09/4






Package artifacts and validateCreate release artifacts9/5


Validate release package



Test release package



Scala release process







Begin Apache votingStart vote on dev@9/7


Start vote on general@







Finalizing and posting the releaseCreate the final release tag on github



Rename, resign and upload the src tar to final dir



Update the website using tag



Release the official pip package



Release the official docker images



After 24 hrs, validate the packages are uploaded



Draft the offical announce email and review



Send out the email on announce@



Update the apache blog



update the aws blog



send internal announcement







CI status:

Release Note

This patch release includes the following bug fixes:


  • No labels

1 Comment

  1. You don't need to pull the julia docs update into the release. The docs on the website for install will show master's content.