Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

docs(kfp): compatibility matrix with TFX #2893

Merged
merged 10 commits into from
Aug 27, 2021
Merged
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
@@ -0,0 +1,14 @@
+++
title = "Compatibility Matrix"
description = "Kubeflow Pipelines compatibility matrix with related software"
weight = 50
+++

## Kubeflow Pipelines Backend and TFX compatibility

| Kubeflow Pipelines Backend & [TFX](https://www.tensorflow.org/tfx) Version | <=0.30.0 | 1.0.0 | 1.2.0 |
Copy link
Contributor

Choose a reason for hiding this comment

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

Is it possible to add HTML to DOCSY?

Kubeflow Pipelines Backend [TFX](https://www.tensorflow.org/tfx) Version
<=0.30.0 1.0.0 1.2.0
<=1.6 Compatible Not fully compatible[1] Not fully compatible[1]️
>=1.7 Not fully compatible[1] Not fully compatible[1]️ Compatible

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes, that's possible too. Let think take a look

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I decided to not do this, because

  1. the empty block under "Kubeflow Pipelines Backend" also looks a little ugly
  2. maintainability of html is worse than markdown

I ended up writing the top left cell as "TFX \ KFP Backend", which I think also conveys the information very well with minimal overhead. What do you think?

Copy link
Contributor

@zijianjoy zijianjoy Aug 26, 2021

Choose a reason for hiding this comment

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

Sounds good, agree with the points you have mentioned.

One thing I would like to say is that we probably want to put the TFX to each row along with the version.

So it is like:

[KFP Backend] <= 1.5 >= 1.7
TFX <= 0.28.0 Fully Compatible ✅ Metadata UI not compatible
TFX 0.29.0, 0.30.0 Visualizations not compatible Metadata UI not compatible
TFX 1.0.0 Metadata UI not compatible Metadata UI not compatible
TFX >= 1.2.0 Metadata UI not compatible Fully Compatible ✅

I imagine this matrix to expand in the future, so each row might mean different dependencies, so putting dependency name and version closer to each other helps readability. For example the future will look like :

[KFP Backend] <= 1.5 >= 1.7
TFX <= 0.28.0 Fully Compatible ✅ Metadata UI not compatible
TFX 0.29.0, 0.30.0 Visualizations not compatible Metadata UI not compatible
TFX 1.0.0 Metadata UI not compatible Metadata UI not compatible
TFX >= 1.2.0 Metadata UI not compatible Fully Compatible ✅
TFDA 0.X.X compatible compatible
TFMA 0.X.X compatible compatible
TFMA > 1.0.0 compatible compatible
PyTorch 0.X.X compatible compatible
PyTorch >= 1.0.0 compatible Compatible

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thanks for the suggestion!
I was planning to add a different table for each dependency, because as you can see there have to be some paragraphs introducing general information about KFP and this dependency.

Copy link
Contributor

Choose a reason for hiding this comment

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

Sounds good on having different tables, thank you Yuan!

| ---- | ---- | ---- | --- |
| <=1.6 | Compatible | Not fully compatible[1] | Not fully compatible[1]️ |
| >=1.7 | Not fully compatible[1] | Not fully compatible[1] | Compatible |

[1] Kubeflow Pipelines UI integration with TFDV, TFMA visualizations and ML Metadata does not work properly. Because of [a bug in TFX 1.0.0](https://github.com/kubeflow/pipelines/issues/6138#issuecomment-899917056), TFX 1.0.0 does not work with any Kubeflow Pipelines backend versions. Besides UI integration, any version of TFX runs on Kubeflow Pipelines.