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

Components - Added readme for TFX components #3637

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
17 changes: 17 additions & 0 deletions components/tfx/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,17 @@
## TFX components for Kubefow Pipelines

Ark-kun marked this conversation as resolved.
Show resolved Hide resolved
Disclaimer: The TFX components in this directory are unofficial and are maintained by the KFP team not the TFX team.

If you experience any issues in this components please create a new issue in the Kubeflow Pipelines repo and assign it to Ark-kun.

These components were created to allow the users to use TFX components in their KFP pipelines, to be able to mix KFP and TFX components.

If your pipeline uses only TFX components, please use the official [TFX SDK](https://www.tensorflow.org/tfx/guide).
Ark-kun marked this conversation as resolved.
Show resolved Hide resolved

See the [sample pipeline](_samples/TFX_pipeline.ipynb) which showcases most of the components.

Aspects and limitations
* These components use the official TFX container image.
* These components run the executors and component classes of the official TFX components.
* These components do not execute TFX drivers, so they do not log metadata themselves (the metadata is logged by the Metadata Writer service instead). The properties of artifacts are currently not logged.
Ark-kun marked this conversation as resolved.
Show resolved Hide resolved
* These components do not execute TFX launchers, so some features might be limited. For example, it's currently not possible to pass `beam_pipeline_args` which prevents some components from utilizing Google Cloud Dataflow.