Table 3 Hyperparameters used in the OD prediction experiment
From: Investigation of the structure-odor relationship using a Transformer model
Values | Optimal OD prediction setting | |
|---|---|---|
Number of heads | 6, 8, 10, 12 | 8 |
Dimension of a single head | 30, 50 | 30 |
Number of encoder layers | 5, 6, 7, 8 | 7 |
Number of decoder layers | 1, 2 | 2 |
\(\tau\) in contrastive loss | 0.3, 0.7, 1.0 | 0.7 |