Table 2 Burgers’ problem: hyperparameters for the AE and CAE networks
From: Deep convolutional architectures for extrapolative forecasts in time-dependent flow problems
Hyperparameters | MLP AE | Convolutional CAE |
|---|---|---|
Encoder layers | [100, 75] | [8, 32] |
Latent dimension (m) | 10, 25, 50 | 12, 25, 50 |
Activation | relu, swish | relu, swish |
Loss function | MSE | MSE |
Learning rate | \(10^{-3},3 \times 10^{-4}\) | \(10^{-3},3 \times 10^{-4}\) |
Model parameters (\(m = 50\)) | 57,060 | 2018 |