We have not had this use case (SAM-like models are currently the only ones with optional inputs. However, they cannot be processed blockwise, because we have not implemented the required coordinate system transforms for its prompt input)
I'm sure a suitable model will come along though -- and bioimageio.core should be ready for it!
We have not had this use case (SAM-like models are currently the only ones with optional inputs. However, they cannot be processed blockwise, because we have not implemented the required coordinate system transforms for its prompt input)
I'm sure a suitable model will come along though -- and bioimageio.core should be ready for it!