Cannot view a vertex AI "dataset" created by service account

[Dhawal Parkar]

Hi team,

I am using a service account via python api to import a csv into cloud storage and then create a vertex ai dataset using the cloud storage as source.

While I can see the csv in cloud storage, I cannot see the vertex ai dataset in the UI.

The service account was created by me and there are no errors when uploading the dataset to the cloud storage or when creating the vertex ai dataset.

Appreciate your help !

Thanks,

Dhawal

[Dhawal Parkar]

Scratch that. You can only view the vertex ai dataset after you have started the training job.

[William Tung]

I'm having a similar occurrence. I cannot see the Vertex AI Dataset nor the Training Job/Model in the UI after creating them with a service account. The training job is complete, but the model is not deployed.

However, I can see these items in the logs, just not the UI.

Process: BigQuery > Vertex AI Dataset > AutoML Tables Training & Model

Did you do anything specific, or did it just appear after you started the training job?

[Dhawal Parkar]

Nothing specific. It appeared immediately after creating the training job. Was the service account created by you ?

[William Tung]

Thanks for the reply

The service account was not created by me personally, it is the Compute Engine default service account associated for the project in my organization. When listing the datasets with automl.TablesClient().list_datasets() all of the datasets created through the API are returned, but none of the ones shown in the Vertex AI > Datasets UI. The only evidence of the API created datasets are in the UI logs.

It is the same GCP project and same compute region, so I'm unsure why API and UI created items are not viewable across each other.

[AlmightySquirrel]

Did you guys solve this issue? I'm facing the same problem. Using API to create dataset, started training as seen in logs but dataset or model etc is not visible in the UI