Code & Data for our COLING 2018 paper "Adaptive Multi-Task Transfer Learning for Chinese Word Segmentation in Medical Text"
Processed open-source datasets are available in data
folder. Annotated medical forum dataset is available in medical_data
folder.
Run python3 main.py -h
to see how to specify your own model settings or datasets.
For training, use the command python3 main.py train -t 2 -p mmd
to train the model with model type 2 and mmd
penalty.
Pls drop me (Junjie Xing) an email at gavinxing9016 at gmail.com if you have any question.