Skip to Main Content
Table 2: 
Hyperparameters in our experiments.
HyperparameterValue
word dropout rate 0.05 
character embedding dimension 128 
CNN window size 
CNN filter number 256 
 
batch size 32 
LSTM hidden size 256 
LSTM dropout rate 0.2 (w/o BERT) 
 0.5 (w/ BERT) 
gradient clipping 5.0 
HyperparameterValue
word dropout rate 0.05 
character embedding dimension 128 
CNN window size 
CNN filter number 256 
 
batch size 32 
LSTM hidden size 256 
LSTM dropout rate 0.2 (w/o BERT) 
 0.5 (w/ BERT) 
gradient clipping 5.0 
Close Modal

or Create an Account

Close Modal
Close Modal