You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently models are selected by the command line option -l which allows adding several models with +. The name and the description "Specify language(s) used for OCR" was historically correct as long as models were available for languages, but meanwhile we have models for scripts, and models can even cover different scripts. In addition it's currently necessary to install all models locally below a common root directory.
A new option -model could support more choices. It could be given more than once which is cleaner than using +, and it could allow qualified model names with an absolute or relative path or a URL for models which are not installed locally.
The text was updated successfully, but these errors were encountered:
Your Feature Request
Currently models are selected by the command line option
-l
which allows adding several models with+
. The name and the description "Specify language(s) used for OCR" was historically correct as long as models were available for languages, but meanwhile we have models for scripts, and models can even cover different scripts. In addition it's currently necessary to install all models locally below a common root directory.A new option
-model
could support more choices. It could be given more than once which is cleaner than using+
, and it could allow qualified model names with an absolute or relative path or a URL for models which are not installed locally.The text was updated successfully, but these errors were encountered: