Table 3 Model training parameters and hyperparameter values.
From: SCBM-Net: a multimodal feature fusion-based dual-channel method for bearing fault diagnosis
Category | Parameter | Value/range |
|---|---|---|
Optimizer | Adam | lr = 0.0001, weight decay = 1e − 5 |
Learning rate schedule | StepLR | Step size = 10, gamma = 0.1 |
Batch size | – | 32 |
Epochs | – | 50 |
Early stopping | Validation accuracy | Patience = 5 |
Loss function | – | CrossEntropyLoss |
Dropout | BiGRU layers | 0.2 |
Random seed | – | 42 |
