Table 6 Hyper-parameters used in our convolutional neural network
From: BioVerbNet: a large semantic-syntactic classification of verbs in biomedicine
Parameters | Values |
|---|---|
Vector dimension | 200 |
Filter sizes | 3,4 and 5 |
Number of filters | 300 |
Dropout probability | 0.5 |
Minibatch size | 50 |
Input size (in tokens) | 500 (documents), 100 (sentences) |