Table 4 Hyperparameter setting
Hyperparameter | Value |
|---|---|
Learning rate | 0.01, 0.001, 0.0001 |
Batch size | 32, 64, 128 |
Epoch | 30, 60, 100 |
Optimizer | Adam |
RoBERTa layer | 12-Layers transformer |
BiLSTM layer | 1, 2, 3 |
GCN layer | 1, 2, 3 |
word embedding dimension | 256, 512, 768 |
Random masking probability | 0.1, 0.2, 0.3 |
Random edge deletion probability | 0.01, 0.05, 0.1 |
Dropout | 0.2 |
Topk | 3 |
Chunk size | 64 |
Chunk overlap | 0.3 |
