This repository contains the official implementation code of the paper Bi-Lingual Named Entity Recognition Using XLM-R and Language Features, accepted at KCC 2022.
Check the datasets. Training sets of English and Korean can be found in this link.
$ pip install -r requirements.txt
Train a XLM-RoBERTa base model
python -m ner_baseline.train_model --train train.txt --dev dev.txt --out_dir . --model_name xlmr_ner --gpus 1 \
--epochs 2 --encoder_model xlm-roberta-base --batch_size 64 --lr 0.0001
Evaluate the trained model
python -m ner_baseline.evaluate --test test.txt --out_dir . --gpus 1 --encoder_model xlm-roberta-base \
--model MODEL_FILE_PATH --prefix xlmr_ner_results
Should you have any questions, feel free to contact fomet1277@naver.com.
The code under this repository is licensed under the Apache 2.0 License.