Table 3 Initial value of the hyperparameters.
From: A new prediction model based on deep learning for pig house environment
Hyperparameter | Initial parameter value |
|---|---|
Number of first convolution layer filter | 10 |
Size of first convolution layer kernel | 16 |
Number of second convolution layer filter | 10 |
Size of second convolution layer kernel | 32 |
Number of neurons in the excitation layer input | 8 |
Number of neurons in the excitation layer output | 32 |
Number of neurons in first GRU | 50 |
Number of neurons in second GRU | 10–100 |
Learning rate | 1e-3–1 |
Batch size | 128 |
Dropout rate | 0.3 |
