Table 4 Parameters of our model in the experiments
From: Multi-task learning for Chinese clinical named entity recognition with external knowledge
Parameters | Value |
|---|---|
Dim of character embedding | 100 |
Dim of radical embedding | 50 |
Number of BiLSTM hidden units | 128 |
Dropout | 0.5 |
Batch size | 32 |
Epochs | 300 |