MIT Press

Table 2:

Hyper-parameter settings of TG-SAN.

Parameter	Value
Word embedding dimension d_e	300
LSTM hidden dimension d_h	150
Dropout rate	0.5
No. of structured representations r	2
Penalization term coefficient λ₁	0.1
Regularization term coefficient λ₂	10⁻⁶
Batch size	64

Parameter	Value
Word embedding dimension d_e	300
LSTM hidden dimension d_h	150
Dropout rate	0.5
No. of structured representations r	2
Penalization term coefficient λ₁	0.1
Regularization term coefficient λ₂	10⁻⁶
Batch size	64