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This repository has been archived by the owner on Nov 17, 2023.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

[HOLD] 1.3.0 release website updates #12509

Merged
merged 3 commits into from
Sep 13, 2018
Merged

Conversation

aaronmarkham
Copy link
Contributor

@aaronmarkham aaronmarkham commented Sep 10, 2018

Description

This PR updates the website for the 1.3.0 release.
Please do not merge until after the release.

After merging this PR:

  1. Merge added new pip packages image for 1.3.0 dmlc/web-data#119 (updates the pip packages graphic)
  2. Update Jenkins CI restricted website build job
    • to have v1.3.0 listed in the tags_to_build field
    • to have 1.3.0 is the tag_list (for what to display on the site)
  3. Enable both the build and the publish jobs (turned off right now due to the tvm issue with the 1.2.0 branch)

Preview

http://34.201.8.176/versions/1.3.0_release/

http://34.201.8.176/versions/1.3.0_release/install/index.html?platform=Linux&language=Python&processor=CPU

http://34.201.8.176/versions/1.3.0_release/install/download.html

@Roshrini
Copy link
Member

@szha to keep track

@TaoLv
Copy link
Member

TaoLv commented Sep 11, 2018

Hi @aaronmarkham , thanks for your effort. Given MKL-DNN backend is already GA in 1.3.0 release, if it is possible to promote how to build it from source, especially for CPU platform? Currently, I didn't find any content about that on the install from source webpage for either 1.2.1 or master.

@lupesko
Copy link
Contributor

lupesko commented Sep 11, 2018

Thanks @aaronmarkham !

@TaoLv I do not think MKL-DNN is GA in 1.3.0. I am aware of ongoing work on master to complete test coverage. Regardless, I think this is a good suggestion to add an explanation for that on the install web page. Would be great if you can contribute this!

@lupesko
Copy link
Contributor

lupesko commented Sep 11, 2018

@aaronmarkham code change looks good, but since this involves UI updates, would be great if you have a host that runs the code to share. Thanks!

@TaoLv
Copy link
Member

TaoLv commented Sep 11, 2018

Thanks for your clarification, @lupesko. I read the message about MKL-DNN GA from here: https://cwiki.apache.org/confluence/display/MXNET/Apache+MXNet+%28incubating%29+1.3.0+Release+Status
I think this decision was disscussed on dev@ and can reflect the status of MKL-DNN backend as well as those effort from the community.

Regarding the doc change, can you kindly point out which source file maps to the install web page?

@aaronmarkham
Copy link
Contributor Author

@lupesko - I added preview links to the description. The one thing that isn't showing is the pip table. I could add a new image and change the links to point to that if we want to see the real preview of that too...

@lupesko
Copy link
Contributor

lupesko commented Sep 12, 2018

Thanks @aaronmarkham - looks good.
I'm wondering if it may be desired to replace/refresh the "Introducing the Scala Inference API" box from 1.2 with something new released with MXNet 1.3.0? Options include Clojure API, ONNX export, TensorRT. WDYT?

Copy link
Contributor

@lupesko lupesko left a comment

Choose a reason for hiding this comment

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

Code changes look good.
Also checked the host running the code: homepage and install page - both looks good.

@aaronmarkham
Copy link
Contributor Author

@lupesko what if we point to the medium post and mention all three...

@Roshrini
Copy link
Member

Is this PR ready to go now?

@aaronmarkham
Copy link
Contributor Author

@Roshrini - yes, the latest preview is up with new news: http://34.201.8.176/versions/1.3.0_release/

@nswamy nswamy merged commit c684909 into apache:master Sep 13, 2018
anirudh2290 pushed a commit to anirudh2290/mxnet that referenced this pull request Sep 19, 2018
* updated for 1.3.0 release

* pip package table update; using picture

* updated news and pip table
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants