An Entity Relation Extraction Method for Few-Shot Learning on the Food Health and Safety Domain

Table 3

Hyperparameter setting.

Hyperparameter	Search space
Bert encoder	Base-uncased; BERT-wmm
Bi-LSTM layer	Number of recurrent layers: [2–14, 16, 17]
Batch size	[128, 256, 512, 1024]
Learning rate	[1 × e^−3, 1 × e^−4, 1 × e^−5]
Epochs	[50, 100, 200]
Loss function	Binary cross-entropy; Negative log-likelihood
Update strategy	[SGD, Adam]