Calamari models?

[bertsky]

The readme and report mention models for Calamari have been trained as well, but I cannot find any.

Note: in the past I have trained both for Calamari 1.x and 2.x because of the OCR-D wrapper being stuck at 1.x. However, there was always a workflow for 2.x as well: by exporting file pairs via ocrd-segment-extract-lines, then using calamari-predict standalone on them, and reintegrating into PAGE via ocrd-segment-replace-text. But the good news is that we do have a working wrapper for 2.x, and with the release of Calamari 2.3, installation problems in a current software environment should also be gone.

So please reconsider Calamari as a valid 3rd option.