|
Algorithm 1 BiLTSM Training Process |
-
1:
Load the training and validation data samples.
-
2:
Initialize model parameters such as minibatch size, maximum epochs, learning rate.
-
3:
Train the model network accordingly and calculate the accuracy error by (14).
-
4:
Adam optimization algorithm is used to compute the corrective parameter and to search for the optimal solution with update of the parameters.
-
5:
Result: Trained model.
-
6:
Save the model.
|