Table 10 Training Hyperparameters Used in Sensitivity Experiments
From: Phenotypic feature-based identification of tea geographical origin using lightweight deep learning
Hyperparameter | Values |
|---|---|
Batch size | 30 |
Optimizer | AdamW |
Weight decay | 0.0001 |
Number of epochs | 100, 200, 300 |
Learning rate | 0.01, 0.001 |
