Closed cneud closed 3 years ago
I had a look on other Makefiles to check for consistency:
So I guess it's up to us how to name this. (There is resource management coming in https://github.com/OCR-D/core/pull/559 but that only pertains to OCR-D.)
I think I checked eynollah for a reference Makefile convention to use.
Moreover, the current version of sbb_binarization
does indeed leverage multiple models - btw this it seems is also not properly exposed here.
Anyway I understand that meanwhile @vahidrezanezhad has trained a newer and uniform model so that future versions will only require a single model file...
Multiple models are indeed used plus this makes naming consistent with other Makefiles.