Generate descriptions from images using masks
Discussions about the Inference Providers feature on the Hub