allenai / bilm-tf

Tensorflow implementation of contextualized word representations from bi-directional language models
Apache License 2.0
1.62k stars 451 forks source link

Contributed ELMo model for Basque #214

Open stefan-it opened 5 years ago

stefan-it commented 5 years ago

Hi,

as part of my research on under-represented languages in NLP, I trained an ELMo model for Basque on various corpora (recent Wikipedia dump, OPUS and Leipzig Corpora Collection).

The repository with link to the model weights and option file can be found here. I also evaluated the ELMo model on downstream tasks like NER and PoS tagging (more to come).

It would be awesome if you can add the model to the contributed elmo models site (https://allennlp.org/elmo). I'm definitely planning to train further ELMo models for more under-represented languages in future :)

Thanks,

Stefan

matt-peters commented 5 years ago

@stefan-it - I have a PR open here to add a link for the Basque model: allenai/allennlp-website#134

Would you please check the contributor information and let me know if you'd like to add any notes or links? I linked to your homepage (from your github profile), but it's easy to change. Thanks for your patience.