You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi @davidxia
Thank you for reaching out. We are currently working on this and will provide an update as soon as we have any new information.
Thank you!
@janasangeetha is there any update on this? If you need any help testing etc., please let me know. TFX 1.16 is working nicely with Kubeflow Pipelines v2.14 with some minor enhancements to logging artifacts.
I'm able to get about 380 tests passing for the components on 3.12.4 with tensorflow 2.16.2 and beam 2.71.0. The same thing with 3.10.14 is passing about 385 tests. If anybody wants to work on this with me, let me know. My goal is to get tfx working on a more modern python and recent version of kubeflow pipelines.
Hi @davidxia@pritamdodeja
Thank you so much for showing interest! We truly appreciate your enthusiasm and support.
Please feel free to fork the repository and submit pull requests if you have any suggestions for enhancements, problem fixes, or ideas. We’d be more than happy to review and collaborate! If you would like to contribute , please follow our CONTRIBUTION GUIDELINES here and submit a pull request.
Thank you again for your interest, and we look forward to seeing your contributions!
Had a go at this in #7851. I am sticking to 3.11 only and keeping the current TF 2.17 pin, so it's pretty minimal. Basically just env markers in the constraint files for the cases where the cp311 wheels want newer apache-beam / protobuf / pyarrow / tensorflow-io-gcs-filesystem than the cp310 ones. 3.9 and 3.10 are untouched.
Also had to bump grpcio-status (the old pin breaks google-api-core on 3.11+) and patch two test files where Python 3.11 changed some mock-library behaviour. All 12 CI jobs green on my fork.
Not touching 3.12 because the cp312 wheels for tfdv/tfx-bsl/ml-metadata aren't on PyPI yet.
Activity
janasangeetha commentedon Feb 18, 2025
Hi @davidxia
Thank you for reaching out. We are currently working on this and will provide an update as soon as we have any new information.
Thank you!
pritamdodeja commentedon Nov 8, 2025
@janasangeetha is there any update on this? If you need any help testing etc., please let me know. TFX 1.16 is working nicely with Kubeflow Pipelines v2.14 with some minor enhancements to logging artifacts.
pritamdodeja commentedon Jan 25, 2026
I'm able to get about 380 tests passing for the components on 3.12.4 with tensorflow 2.16.2 and beam 2.71.0. The same thing with 3.10.14 is passing about 385 tests. If anybody wants to work on this with me, let me know. My goal is to get tfx working on a more modern python and recent version of kubeflow pipelines.
Pritam
bharatjetti commentedon Jan 28, 2026
Hi @davidxia @pritamdodeja
Thank you so much for showing interest! We truly appreciate your enthusiasm and support.
Please feel free to fork the repository and submit pull requests if you have any suggestions for enhancements, problem fixes, or ideas. We’d be more than happy to review and collaborate! If you would like to contribute , please follow our CONTRIBUTION GUIDELINES here and submit a pull request.
Thank you again for your interest, and we look forward to seeing your contributions!
Add support for Python 3.11
nblomerus commentedon May 18, 2026
Had a go at this in #7851. I am sticking to 3.11 only and keeping the current TF 2.17 pin, so it's pretty minimal. Basically just env markers in the constraint files for the cases where the cp311 wheels want newer apache-beam / protobuf / pyarrow / tensorflow-io-gcs-filesystem than the cp310 ones. 3.9 and 3.10 are untouched.
Also had to bump grpcio-status (the old pin breaks google-api-core on 3.11+) and patch two test files where Python 3.11 changed some mock-library behaviour. All 12 CI jobs green on my fork.
Not touching 3.12 because the cp312 wheels for tfdv/tfx-bsl/ml-metadata aren't on PyPI yet.
Any feedback would be appreciated. Thank you.