Google Cloud AI Platform: Image Data Labeling Service Error - Image URI not valid

Question

Error in Google Cloud Data labeling Service:

• I am trying to create a dataset of images in Google's Data labeling service.
• Using a single image to test it out.
• Created a Google storage bucket named: my-bucket
• Uploaded an image to my-bucket - image file name: testcat.png
• Created and uploaded a csv file (UTF-8) with URI path of image stored inside it.
• image URI path as stored in csv file: gs://my-bucket//testcat.png
• Named the csv file : testimage.csv
• Uploaded the csv file in the gs bucket - my-bucket.
• i.e. testimage.csv, and testcat.png are in the same google storage bucket (my-bucket).

When I try to create the datasset in google console, GCP gives me the following error message:

** Failed to import dataset gs://my-bucket/testcat.png is not a valid youtube uri nor a readable file path.**

I've checked multiple times and the URI for this image in Google is exactly the same as what I've used. I've tried at least 10-15 times ... the error persists.

Any one faced and successfully resolved this issue? Your help is greatly appreciated.

Thanks!

Answer 1

As you can see in our AI Platform Data Labeling Service documentation, there is a service update due to the coronavirus (COVID-19) health emergency that states that data labeling services are limited or unavailable until further notice.

• You can't start new data labeling tasks through the Cloud Console, Google Cloud SDK, or the API
• You can request data labeling tasks only through email at cloudml-data-customer@google.com
• New data labeling tasks can't contain personally identifiable information