Been deep diving into visual model architectures. Currently running some small evaluations on an idea around unlabeled mask segmentation that will run efficient

Been deep diving into visual model architectures. Currently running some small evaluations on an idea around unlabeled mask segmentation that will run efficiently on mobile phone hardware. Looking promising so far!