TFX 0.25.0 is Out

[Dhruvesh Talati]

Dear TFX community,

We are pleased to announce that TFX 0.25.0 is released. Please feel free to test this and report any issues you find in the process through GitHub.

The GitHub Url: https://github.com/tensorflow/tfx

Some highlighted features included in this version:

• Added CloudTuner KFP e2e example running on Google Cloud Platform with distributed tuning.
• Introduced Kubeflow V2 DAG runner, which is based on Kuberflow IR spec. Same as KubeflowDagRunner it will compile the DSL pipeline into a payload but not trigger the execution locally.
• TFX will also be hosting nightly packages at https://pypi-nightly.tensorflow.org on Google Cloud. To install the nightly package use the following command: pip install -i https://pypi-nightly.tensorflow.org/simple tfx

And much more!

Full release notes: https://github.com/tensorflow/tfx/releases/tag/v0.25.0

PyPI wheels: TFX 0.25.0

Docker Image: tensorflow/tfx:0.25.0

Full release notes for a few notable dependencies of TFX:

TensorFlow Serving (TFS): https://github.com/tensorflow/serving/releases/tag/2.3.0

TensorFlow Data Validation (TFDV): https://github.com/tensorflow/data-validation/releases/tag/v0.25.0

TensorFlow Transform (TFT): https://github.com/tensorflow/transform/releases/tag/v0.25.0

TensorFlow Model Analysis (TFMA): https://github.com/tensorflow/model-analysis/releases/tag/v0.25.0

TFX Basic Shared Libraries (TFX_BSL): https://github.com/tensorflow/tfx-bsl/releases/tag/v0.25.0

ML Metadata (MLMD): https://github.com/google/ml-metadata/releases/tag/v0.25.0

For more information

To learn more about TFX check out the TFX website, dive into other posts in the TFX blog, watch our TFX playlist onYouTube, and subscribe to the TensorFlow channel.

--

Cheers,

Dhruvesh Talati, on behalf of TFX::OSS team