Execute the following command to run the BERT model for Aspect Extraction (NER) task:
!python src/run_ae.py --data_dir "ae/laptop" --eval_data_dir "ae/rest" --output_dir "pt_model/laptop_ae" --do_train --num_train_epochs 10 --do_valid
Execute the following command to run the BERT model for Aspect Sentiment Classification task:
!python src/run_asc.py --data_dir "asc/laptop" --eval_data_dir "asc/rest" --output_dir "pt_model/laptop_asc" --do_train --num_train_epochs 10 --do_eval --do_valid
More tested approaches can be found in the Methods
directory, the structure is as follows
── Methods
├── Attention-Based LSTM
├── SVM
└── Word2Vec