Table 4 Parameters of model experiments
Parameters | Value |
|---|---|
Epoch | 200 |
Learning rate | 1e−3 |
Batch size | 256 |
Weight decay | 5e−4 |
Loss function | BCELoss |
Drug embedding dimension | 64 |
Drug length | 100 |
Number of attention heads | 5 |
Number of Graph Attention Convolution layers | 3 |
Convolutional kernel size | [4,6,8] |