/JBLACN

source code and dataset for JBLACN

Primary LanguagePythonApache License 2.0Apache-2.0

Requirement

  • Python3.6
  • PyTorch: 1.0.1

Train models

  • Download data and word embedding
  • Run the script:
nohup python -u joint_main.py --use_pre_trained_model False --use_crf true 
--hidden_dim 300 --attention_query_input_size 303 --num_attention_head 5 
--iteration 1000 --learning_rate 0.02 >log.txt 2>&1 &

Acknowledgments

NCRF++