How to see which datarows have MAL pre-labels attached - Python SDK

Python SDK
, ,
1

Hello, I’ve used the model assisted labeling pipeline to attach pre-labels to a data batch. If i want to create a pipeline that finds all images across all projects that don’t have pre-labels, how would I do so?

I am currently iterating through all projects, and exporting the queued data rows. Ideally, a data row would not show up in queued if it already has a MAL pre-label:

project = client.get_project('clfqxx1vb05ln07wq9ucu8fh6')
data_rows = project.export_queued_data_rows()
# insert some logic to filter out data rows that already have pre-labels attached

thanks in advance!

2

Hey Chris, thanks for your question!

1/ we’re working on making MAL visualisable in your labeling projects, so it’s easier to identify visually which data rows have/don’t have MAL

2/ it’s on the roadmap to have filters such as “has a pre-labels”/“doesn’t have a pre-label”

Meanwhile, the solution you describe is the correct one :+1: