search

Alibaba-NLP / ACE

[ACL-IJCNLP 2021] Automated Concatenation of Embeddings for Structured Prediction
Other
298 stars 44 forks source link

Using your best model for inference for NER in production #45

Closed Dimiftb closed 2 years ago

Dimiftb commented 2 years ago

Hi there,

Is there an easy way to use your pre-trained model for inference? If not, do you have plans perhaps to release it in huggingface or any other place so that we can use it in production as you still hold the SOTA on CoNLl dataset?

Thanks

wangxinyu0922 commented 2 years ago

Hi, you can follow the guide in named_entity_recognition.md for inference. I have no plan to release it in huggingface currently.

Dimiftb commented 2 years ago

Hi @wangxinyu0922

Appreciate your response. I'm struggling to find the particular information.

Let's say I've got a sentence "Berlin is a great city". How would I go about running the model on it so that I can extract the named entities.

Could you please point me to the exact location in the readme file where I can read about how to achieve that if it's there?

Thank you.

wangxinyu0922 commented 2 years ago

Check this part first, which guides the predict file formatting and how to use the model for predictions.

Dimiftb commented 2 years ago

Appreciate your help.