Table 1 Hyperparameter configuration.
From: Multiscale wavelet attention convolutional network for facial expression recognition
Hyperparameter | Value |
|---|---|
Learning rate | \(\:\eta\:=0.0001\) |
Training process (TP) | \(\:5\) |
Max epochs | \(\:300\) |
Batch size | \(\:64\) |
Dropout | \(\:dropout\_rate1=0.5\:\text{a}\text{n}\text{d}\:dropout\_rate2=0.3\) |
Batch normalization | \(\:Batch\_Norm1=192\:\text{a}\text{n}\text{d}\:Batch\_Norm2=256\) |
Patience | \(\:30\) |
Delta | \(\:0.00001\) |
Optimizer | Adam Optimizer27 |
