Hi all,
I will highly appreciate to get some valuable inputs from the community over below:
I have created an label studio instance, setup a project, and created a ml-backend for performing auto labelling. (Using docker setup) However, when I upload a new image to the project, the ‘setup’ function is called but not the predict function. I have validated that auto-labelling is set true. And if I open an uploaded image, than the ‘predict method’ is called. And If I annotate manually: then the ‘fit’ method is called.
I have read the label studio ml documentations, and the ‘predict’ function should have been triggered on a new image upload.
I will appreciate any help on: ‘Reasoning behind the ‘predict’ function not being called on a new image upload itself’.
Thanks.