Oswaldo Gomez comments

Results 20 comments of


                                            Oswaldo Gomez

Unable to get results from sample query

Another comment is that, this seems to be using version 1.

Deployment to K8s fails on Release pipeline in "./aml_service/51-deployOnAks.py"

Same here!

Add support for environment variables at pyproject.toml

This would be incredibly convenient for CI/CD pipelines and for reusing already available private repository urls in the form of environment variables.

[backend] mysql 8.0 support (upstream ml-metadata issue)

I'm confused by these two things at the same time, from the title of this issue - we still depend on google/ml-metadata, which does not fully support MySQL 8.0. In...

Facing KeyError: 0 when trying to load model from S3 URI

I have the exact same issue. Let me know if you found a workaround.

update protobuf version

This MR would be incredible to merge as we cannot install `kserve` with latest version of KFP in the same package. Are there any plans to revive this MR? Thanks...

[sdk] Unable to aggregate results over ParallelFor in Kubeflow V2 using V1 workarounds such as `.after()`

We are also impacted by this. We see this as a critical blocker for us since we have more and more tasks that we want to run in parallel. Specially...

[sdk] Unable to aggregate results over ParallelFor in Kubeflow V2 using V1 workarounds such as `.after()`

I just tested this on Kubeflow 1.8.1 and using KFP SDK 2.5.0 and 2.8.0, but none work. Here's the sample dummy pipeline ``` from kfp import compiler, dsl from typing...

[sdk] Unable to aggregate results over ParallelFor in Kubeflow V2 using V1 workarounds such as `.after()`

@connor-mccarthy could we reopen this issue based on the comments above?

[sdk] Unable to aggregate results over ParallelFor in Kubeflow V2 using V1 workarounds such as `.after()`

We have many new ML use cases that we are onboarding and we are forced to do so in V1 pipeline because of this bug fix. Thanks everybody for the...