Multitask Learning for Aspect-Based Sentiment Classification

Table 3

Hyperparameters used in the experiment.

Parameter	Value
Dropout rate	0.5
Batch size	32
Learning rate	2e−5
Max epoch	20
Max sequence length	128
Optimizer	Adam