BioBERT for NER

To train an NER model with BioBERT-v1.1 (base), run the command below.

Before running this, make sure you have generated the pre-processed dataset using the generate_data.py file with the command mentioned in the parent directory.

Additional Requirements

seqeval: Used for NER evaluation (pip install seqeval)

Training

export SAVE_DIR=./output
export DATA_DIR=./dataset

export MAX_LENGTH=128
export BATCH_SIZE=16
export NUM_EPOCHS=5
export SAVE_STEPS=1000
export SEED=0

python run_ner.py \
    --data_dir ${DATA_DIR}/ \
    --labels ${DATA_DIR}/labels.txt \
    --model_name_or_path dmis-lab/biobert-large-cased-v1.1 \
    --output_dir ${SAVE_DIR}/ \
    --max_seq_length ${MAX_LENGTH} \
    --num_train_epochs ${NUM_EPOCHS} \
    --per_device_train_batch_size ${BATCH_SIZE} \
    --save_steps ${SAVE_STEPS} \
    --seed ${SEED} \
    --do_train \
    --do_eval \
    --do_predict \
    --overwrite_output_dir

Results

	precision	recall	f1-score
ADE	0.6351	0.5680	0.5997
Dosage	0.9254	0.9485	0.9368
Drug	0.9580	0.9542	0.9561
Duration	0.8119	0.9021	0.8546
Form	0.9546	0.9456	0.9501
Frequency	0.9707	0.9668	0.9688
Reason	0.7203	0.7348	0.7275
Route	0.9530	0.9525	0.9527
Strength	0.9807	0.9846	0.9827
micro avg	0.9327	0.9330	0.9328
macro avg	0.9253	0.9225	0.9230