Skip to main content

Table 2 The runtime of the experiments

From: Improving the recall of biomedical named entity recognition with label re-correction and knowledge distillation

Models

Time (min)

Weakly labeled (BiLSTM-CRF)

320

Training (BiLSTM-CRF)

2

Distillation (BiLSTM-CRF)

625

Weakly labeled (BioBERT-CRF)

334

Training (BioBERT-CRF)

4

Distillation (BioBERT-CRF)

554

  1. “Time” denotes training time for one epoch. “Weakly labeled” and “Training” are the training time of the model trained on the weakly labeled dataset and CDR training dataset, respectively. “Distillation” is the training time of knowledge distillation