OpenNMT / CTranslate

Lightweight C++ translator for OpenNMT Torch models (deprecated)
https://opennmt.net/
MIT License
79 stars 50 forks source link

Added command line option to tag unknown words in output #48

Closed fdalvi closed 5 years ago

fdalvi commented 5 years ago

A new command line option --replace_unk_tagged has been added to mirror the functionality in OpenNMT (http://opennmt.net/OpenNMT/translation/unknowns/). The option works just like --replace_unk, except that it wraps the unknown word in a ⦅unk:xxxxx⦆tag. This is useful for post-processing the unknown tokens in the output.

guillaumekln commented 5 years ago

Thanks!