Kubeflow pipeline artifact visualization

[Dhaval08]

What steps did you take:

I am trying to visualize artifacts for Pipeline runs.
Following is what I write to mlpipeline-ui-metadata.json

metadata = {
    'outputs': [{
        'type': 'markdown',
        'source': 'artifact.md',
    }]
}

I read that the supported source locations are http, https, Amazon S3, Minio, and Google Cloud Storage. Is there any way to use local file system as source for artifact visualization? If not, is there any work around to this? My training code simply writes certain files (like markdown or tensorboard logs) to local file system and I would like to visualize them on Kubeflow dashboard.

What happened:

Artifact tab is not loading any artifact

[Ark-kun]

Please try the following:

metadata = {
    'outputs': [{
        'type': 'markdown',
        'source': 'your *markdown* here',
        'storage': 'inline',
    }]
}

P.S. @Bobgy I think resolving the following issues will make this significantly easier for the users:
#3970
#3510

[Dhaval08]

@Ark-kun
Thank you.
I was trying to give the local filesystem path as the source. Giving as S3 path worked for me. Any update on supporting local path as source for these artifacts?

Also, does Kubeflow support real time visualization of tensorboard logs? I did try out the steps from documentation. However, the visualization seems to get created only after the job has finished its execution.
Thank you.

[Bobgy]

There cannot be support for local. Do you mean volume?

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically closed because it has not had recent activity. Please comment "/reopen" to reopen it.