Allow a configuration for inferencing of SFM files that may differ from project to project. #306
Labels
enhancement
New feature or request
pipeline 6: infer
Issue related to using a trained model to translate.
We could save translators some work if our inferencing tools could adapt to individual projects. This has been requested by several project teams.
Can we provide a way to specify for each marker whether its contents whether it should be Translated, Deleted, or Passed-through unchanged? Perhaps an inferencing section in the config file, or a separate inferencing config file would work well.
There's some complexity to this since some markers are line-level markers and some are inline. If we specify that a line-level marker should be omitted, then what should happen to in-line markers within that line that we ask to be translated.
There may also be formatting marker that affect only the display of data, we don't really have the option of passing those through or 'translating' them.
The text was updated successfully, but these errors were encountered: