Table 3 Bayesian Optimization for Hyperparameters in ABT-MPNN
Hyperparameters | Values |
|---|---|
Message-passing iteration (T) | 2, 3, 4, 5, 6 |
Inter-atomic feature scaler (λ) | [0, 0.5] (Interval: 0.05) |
Hidden dimension (d) | [300, 2400] (Interval: 100) |
Dropout probability (p) | [0, 0.4] (Interval: 0.05) |