Fix tensorboard.py #9743

[okwrtdsh]

Description

(Brief description on what this PR is about)
Fixed TensorBoard summary writer.
See #9743.

Checklist

Essentials

Please feel free to remove inapplicable items for your PR.

• The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant JIRA issue created (except PRs with tiny changes)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage:
• Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
• Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
• Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
• Code is well-documented:
• For user-facing API changes, API doc string has been updated.
• For new C++ functions in header files, their functionalities and arguments are documented.
• For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
• Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$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

• tensorboard.SummaryWriter to mxboard.SummaryWriter

Comments

tensorboard.py does not work any more, because new tensorboard does not have a Class named SummaryWrite

[rahul003]

I think we should be using mxboard instead of forcing users to install Tensorflow to use MXNet with tensorboard. You can import SummaryWriter from mxboard with the same API as before.

[szha]

@rahul003 this is a reasonable maintenance change for contrib feature.

[rahul003]

It comes off odd that we would want users to install Tensorflow to do something in MXNet. If it was some other library that would be a different story. Why not use a library MXNet community has built (MXBoard) for this specific purpose of using Tensorboard with MXNet?

In case I got it wrong, does installing Tensorboard now require installing Tensorflow? In that case, there's no way around it then. EDIT: That's not the case, just verified.

[okwrtdsh]

Thanks for the comments. I've changed Tensorflow to MXBoard.