search

yzhangcs / parser

:rocket: State-of-the-art parsers for natural language.
https://parser.yzhang.site/
MIT License
825 stars 138 forks source link

Feature Request: Customizing Stanza Pipeline #92

Closed nu11us closed 2 years ago

nu11us commented 2 years ago

Is it possible to add the ability to specify parameters in the stanza tokenization pipeline to the API?

I'd like to do some testing using SuPar with a pretrained Stanza tokenizer like so using the tokenize_model_path argument.

yzhangcs commented 2 years ago

@nu11us Hi, currently SuPar does not expose the interface of passing args to stanza tokenizer, you can manually modify the code here if you wish. https://github.com/yzhangcs/parser/blob/33373037ef982e5bbda1c375f29f105f79ec41f5/supar/utils/tokenizer.py#L9-L12

nu11us commented 2 years ago

Thank's for the advice, I'll make a fork for that and try it out.