I have an image where I have to do a multi-label classification and additionally count the number of a specific item in each image. I'm trying to setup a labeling task so I can enter a continuous number (0-100 for example), but there doesn't seem to be support for it.
Additionally, does the labeling have capabilities to pre-choose a "default" value?
Does anyone have an idea?
Hello,
As you have rightly mentioned, Vertex AI does not presently support object counting in an image. Please feel free to reach out to submit a feature request via the issue tracker link[0] to the Vertex AI product team about such implementations.
As mentioned in this article[1], there are only three ways to assign labels to your training data items:
-- Add the data items to your dataset with their labels already assigned, for example using a commercially available dataset
-- Assign labels to the data items using the Cloud Console
-- Request to have human labelers add labels to the data items
At this time, none of these options provide ways to pre-choose a "default" value. May be there may be workarounds to explore using your own human labelers via your instructions.
[0]https://developers.google.com/issue-tracker/#public_users
[1]https://cloud.google.com/vertex-ai/docs/datasets/data-labeling-job
User | Count |
---|---|
2 | |
1 | |
1 | |
1 | |
1 |