Table 3 Hyper parameters of the residual neural network
From: Attention-based deep residual learning network for entity relation extraction in Chinese EMRs
Parameter | Description | Value |
|---|---|---|
d w | Dimension of word embedding | 100 |
d p | Dimension of position embedding | 5 |
k | Window size | 3 |
m | Number of filters | 128 |
B | Batch size | 50 |
λ | Learning rate | 0.01 |
p | The ration of dropout | 0.3 |