- Synonym dataset
datasets/synonyms/*is built on Chinese Synonym Dataset: 同义词词林. - Pre-train word embedding:
You can use datasets/synonyms/* or dataset else you built.
Download from the above Pre-train word embedding.
You can install dependencies by:
pip install -r requirements.txtpython main.py --train datasets/synonyms/train \
--dev datasets/synonyms/dev \
--test datasets/synonyms/test \
--embedding /path/to/embedding_file \
--outputs /path/to/outputs_dir@Apache 2.0 (Except for datasets)